Huge news: Clarifai has agreed to license our AI inference & compute orchestration IP — plus the patent portfolio behind it – to @nebiusai.
Our core team is joining them to keep building.
Our technology becomes a key part of Nebius Token Factory, the inference platform inside their full-stack AI cloud. Faster inference. Bigger scale.
Can't wait to ride this rocket ship and build the foundation for the next decade of AI inference. 🚀
Clarifai
Clarifai Inc. is an artificial intelligence (AI) company that specializes in computer vision and uses machine learning and deep neural networks to identify and analyze images and videos. More on Wikipedia
-
Clarifai reposted
Matt Zeiler
@mattzeiler
-
Clarifai
@clarifai
NVIDIA Nemotron 3 Nano Omni is now available on Clarifai with Zero Day support.
A 30B A3B multimodal reasoning model built for agent workflows across documents, images, video, audio, and text.
Why it stands out:
• Multimodal input across text, image, video, and audio
• Hybrid MoE + Transformer-Mamba architecture
• 300K context window
• Runs on a single H100, H200, or B200
• 400 tokens/sec on Clarifai Reasoning Engine
Clarifai
@clarifai
Try the new model here: clarifai.com/nvidia/chat-completion/models/nemotron-nano-…
Clarifai
@clarifai
Learn more about Nemotron 3 Nano Omni here: www.clarifai.com/blog/nvidia-nemotron-3-nano-omni-on-clar…
-
Clarifai
@clarifai
New SotA open source model, new pole position for Clarifai Reasoning Engine.
Kimi K2.6 cooking at 164 tokens per second.
Clarifai
@clarifai
Try it now in our playground: clarifai.com/playground?model=Kimi-K2_6&user_id=moonshota…
-
Clarifai
@clarifai
Qwen3.6-35B-A3B is now live on Clarifai. 🚀
The model delivers frontier-level agentic coding performance with a MoE architecture that only activates 3B out of 35B total parameters per token.
Apache 2.0 licensed and runs locally.
Key capabilities:
- 73.4% on SWE-Bench Verified for real-world GitHub issue resolution
- 51.5% on Terminal-Bench 2.0 (highest among open models)
- 37.0% on MCPMark for tool use and function calling
- Multimodal support (text, image, video)
- 262K native context, extendable to 1M tokens
- Thinking Preservation: retains reasoning traces across multi-turn conversations for better agent consistency
Access via API or deploy on your own dedicated compute.
Clarifai
@clarifai
Try the model here: clarifai.com/qwen/qwen-VL/models/Qwen3_6-35B-A3B-FP8
-
Clarifai
@clarifai
Clarifai 12.3: Introducing KV Cache-Aware Routing! 🚀
LLM inference with multiple replicas typically uses random load balancing. But replicas aren't interchangeable. Each builds up different KV cache state from the requests it processes.
When requests land on replicas without relevant context cached, models recompute everything from scratch. This wastes GPU cycles and increases latency.
KV Cache-Aware Routing automatically detects prompt overlap and routes requests to replicas with relevant cache state already loaded. Shared system prompts are computed once and reused. RAG context is cached when multiple users query similar documents.
Zero configuration required. Deploy a model with multiple replicas and it works.
Also new: Warm Node Pools for faster scaling, Session-Aware Routing to keep user requests on the same replica, Prediction Caching for identical inputs, and Clarifai Skills for AI coding assistants.
Read the full release blog: www.clarifai.com/blog/clarifai-12.3-introducing-kv-cache-… -
Clarifai
@clarifai
Do you feel the need for speed?
Come learn how Clarifai breaks speed records for inference across multiple models at Booth #405 at #humanx.
NVIDIA AI Infrastructure
@NVIDIAAIInfra
⚡ In just 10 days, leading inference providers propelled Kimi K2.5 up the @ArtificialAnlys leaderboard by adopting key elements of NVIDIA’s Inference Reference Architecture.
Incredible work by @basetenco, @clarifai, @DeepInfra, @Eigen_AI_Labs, @FireworksAI_HQ, @friendliai, @LightningAI, @nebiusai, @togethercompute, @wandb by @CoreWeave.
From custom kernel optimizations to NVFP4, disaggregated serving, and speculative decoding, this extreme full-stack co-design of the NVIDIA Blackwell platform is driving major gains in both performance and efficiency.
The result? The lowest token cost at scale.
Get started with your preferred NVIDIA partner or explore our Inference Reference Architecture documentation ➡️ nvda.ws/4twlPIH -
Clarifai
@clarifai
Day 2 at HumanX! 🚀
Clarifai Reasoning Engine is running at booth #405. Custom CUDA kernels, speculative decoding optimized for reasoning workloads, and adaptive optimization.
That's how we deliver 500+ TPS on Kimi K2.5 and lead on Qwen3.5 and other top reasoning models benchmarked by Artificial Analysis.
But software optimization is only half the story. Tomorrow at 11 AM, Alfredo Ramos (our CPTO) is presenting with @Vultr on the infrastructure side: how to design network architecture that scales from edge deployments to hyperscale data centers for AI workloads.
Vultr Theatre Session at booth #825. -
Clarifai
@clarifai
Day 1 at HumanX! 🚀
Clarifai is at booth #405 showing how we're running reasoning models in production.
We hit 500+ TPS on Kimi K2.5 - first provider to cross that threshold - and we're currently leading on Qwen3.5 at 290 TPS on Artificial Analysis benchmarks.
Getting that performance requires more than optimized inference code. You need network infrastructure built for AI workloads from the ground up.
Tomorrow at 11 AM, our VP of Strategy Sajai Krishnan is presenting on exactly that: building AI data centers that can handle edge to hyperscale deployments.
@Vultr Theatre Session at booth #825. -
Clarifai
@clarifai
Vendor lock-in is the biggest AI infrastructure risk nobody's planning for.
The attempted federal ban on Anthropic in late February exposed what happens when you build your entire AI stack on a single vendor. Government agencies and contractors scrambled overnight. The ban was blocked by a federal judge as illegal retaliation, but the lesson stands: you're one policy decision away from a forced migration.
Building on a single API means when that vendor becomes unavailable, you're rebuilding from scratch. Claude today. GPT tomorrow. Gemini if you have to. Your infrastructure should handle that swap without breaking.
That's how Clarifai is built. Deploy any open-source reasoning model on your own infrastructure - Kimi K2.5, Qwen3.5, GLM, or your own fine-tuned models.
We're leading on Kimi K2.5 and Qwen3.5 on Artificial Analysis benchmarks because the infrastructure is purpose-built for reasoning workloads, not locked to a single model.
Build your AI stack so you can swap the engine without rebuilding the car. -
Clarifai
@clarifai
One week until HumanX 2026. 🚀
Clarifai is bringing production-ready AI infrastructure to San Francisco April 6-9.
We're now leading on both Kimi K2.5 (500+ TPS) and Qwen3.5-397B (290 TPS), benchmarked by Artificial Analysis.
Deploy any AI model, on any compute, at any scale. Our platform handles the complexity - from compute optimization to production inference.
See Clarifai Reasoning Engine live at booth #405. Schedule time with the team here.
Clarifai
@clarifai
Book a meeting with the team here: www.clarifai.com/meet-clarifai-at-humanx-2026
-
Clarifai
@clarifai
Clarifai is heading to HumanX 2026 in San Francisco. 🚀
At GTC this week, we announced 414 tokens per second on Kimi K2.5 - first provider to reach this performance. We're also leading on Qwen3.5-397B with 290 TPS, benchmarked by Artificial Analysis.
Now we're bringing that momentum to #humanx April 6-9.
Come see how Clarifai Reasoning Engine delivers production-ready performance for reasoning models and agentic workloads.
Find us at booth #405 or schedule time with Matthew Zeiler, Sajai Krishnan, or Douglas Shapiro.
Book a meeting here: www.clarifai.com/clarifai-at-humanx-2026 -
Clarifai
@clarifai
Day 3 for GTC is here and there is a new pole sitter for Qwen3.5 performance.
Clarifai Reasoning Engine approaches 300 tokens per second - 245% faster than a vanilla install.
Clarifai
@clarifai
Verified on @ArtificialAnlys - check out the performance stats here: artificialanalysis.ai/models/qwen3-5-397b-a17b/providers
Clarifai
@clarifai
If you're at GTC, meet with us to see the speed first hand.
www.clarifai.com/clarifai-at-gtc-nvidia-event-2026 -
Clarifai
@clarifai
Day 2 at GTC. 🚀
Clarifai Reasoning Engine hit 410 tokens per second on Kimi K2.5, making us one of the first providers to break the 400 TPS barrier on a trillion-parameter reasoning model.
We're also running strong benchmarks on MiniMax M2.5, Qwen3.5, and GPT-OSS-120B. Production-ready performance for reasoning workloads. ⚡
Clarifai
@clarifai
Meet the team at GTC here:
www.clarifai.com/clarifai-at-gtc-nvidia-event-2026 -
Clarifai
@clarifai
Jensen breaks down the vital importance of inference architecture in this new economy in this year's GTC keynote.
"Every. Single. Company. Will be thinking about their token factory effectiveness."This media couldn't be displayed. See it on the original post. -
Clarifai
@clarifai
👀
What company broke the 400 token per second barrier on @Kimi_Moonshot K2.5 first?
Check out Clarifai in Jensen's @nvidia GTC keynote.
Clarifai
@clarifai
If you're at the show, we'd love to see what you're building. Book some time with @mattzeiler, Alfredo Ramos or Dr. Michael Gormish here:
www.clarifai.com/clarifai-at-gtc-nvidia-event-2026 -
Clarifai
@clarifai
Clarifai 12.2: Three-Command CLI Workflow for Model Deployment! 🚀
Model deployment shouldn't require juggling multiple tools and configuration files. With Clarifai 12.2, you can go from development to production in three CLI commands:
→ model init – Scaffold with automatic GPU selection → model serve – Test locally with production parity → model deploy – Deploy with automatic infrastructure provisioning
The CLI handles dependency management, GPU selection, and deployment orchestration automatically. Multi-cloud instance discovery works across AWS, Azure, DigitalOcean, and other providers.
Also new in 12.2:
• Training on Pipelines (Public Preview)
• Video Intelligence for real-time stream analysis
• Dynamic nodepool routing and deployment enhancements
Clarifai
@clarifai
Read the full release blog:
www.clarifai.com/blog/clarifai-12.2-three-command-cli-wor… -
Clarifai
@clarifai
Heading to NVIDIA GTC 2026?
As GenAI moves into production, two challenges keep coming up: GPU scarcity and rising inference costs.
Scaling AI is no longer just about choosing the right model. It’s about how efficiently you use the compute behind it.
At GTC, Clarifai is joining @Vultr for a Theater Session to discuss how enterprises are improving throughput, controlling inference spend, and making better use of GPUs across cloud and hybrid environments.
Our VP of Strategy, Sajai Krishnan, will share what teams are learning as they move from pilots to real production workloads.
March 17 | 4:00 PM
Booth #1631
If you’re building or operating AI systems at scale, join the conversation:
www.clarifai.com/clarifai-at-gtc-nvidia-event-2026
#gtc -
Clarifai
@clarifai
Most MCP servers live in local dev environments.
But real applications need them as stable endpoints.
MCP servers are how LLMs connect to tools and external systems. Web search. GitHub. Databases. Internal APIs. They expose capabilities that models can call through structured tool definitions.
The problem is that most MCP servers run as simple stdio processes. That works for experiments, but not for production systems where agents, applications, and teams need reliable access.
The missing step is turning those servers into accessible API endpoints.
In this guide we walk through how to deploy an MCP server so its tools can be discovered and invoked by any LLM that supports function calling. No changes to the server itself. Just package it, deploy it, and expose the tools through a stable endpoint.
Once deployed, your MCP server becomes shared infrastructure for agents and applications.
Full walkthrough here:
www.clarifai.com/blog/how-to-deploy-mcp-servers-as-an-api… -
Clarifai
@clarifai
Choosing an open-source LLM for production is harder than it looks.
Benchmarks are helpful. But they don’t tell you how a model will behave under your workload, your latency targets, or your cost constraints.
Before comparing models, ask:
• What exact task are we solving?
• Where will this run — single GPU, multi-node, real-time, batch?
• What are our non-negotiables — licensing, privacy, predictable cost?
Different workloads require different trade-offs. Multi-step reasoning, structured outputs, long context, code generation — they all stress models differently.
And evaluation should reflect production reality, not leaderboard scores.
Model choice is only half the equation. How you deploy, scale, and optimize that model determines whether it actually works in production.
We broke this down in detail here:
www.clarifai.com/blog/how-to-choose-the-right-open-source… -
Clarifai
@clarifai
Clarifai will be at NVIDIA GTC 2026.
As GenAI moves from pilot to production, two pressures are hitting hard:
Rising inference costs.
GPU scarcity across environments.
Scaling AI is not about bigger models alone. It’s about orchestration and making better use of the compute you already have.
At GTC, we’re joining @Vultr for a live Theater Session to discuss what this looks like in practice.
Our VP of Strategy, Sajai Krishnan, will share how enterprises are improving throughput, lowering compute spend, and gaining greater control over production AI systems without locking into a single accelerator ecosystem.
Tuesday, March 17 | 4:00 PM
Booth #1631
If you’re building or operating AI at scale, join us:
www.clarifai.com/clarifai-at-gtc-nvidia-event-2026
#gtc2026 -
Clarifai
@clarifai
MCP servers are easy to run locally.
But running them reliably as shared infrastructure is a different problem.
We wrote a guide on how to deploy any stdio-based MCP server as a managed endpoint on Clarifai.
Just define the MCP command, deploy it, and access its tools through a stable endpoint from any LLM that supports tool calling.
We used the DuckDuckGo MCP server as an example, but this works for any MCP server.
Full guide: www.clarifai.com/blog/how-to-deploy-mcp-servers-as-an-api… -
Clarifai
@clarifai
Everyone is building agentic AI.
Very few are running it in production.
The model performs.
The demo impresses.
Then reality shows up.
Agents do not just generate text. They plan across steps. Call tools. Hit APIs. Maintain state. Run workloads that may take minutes or hours. And when they fail, they fail mid-process.
That is not a model problem.
That is a systems problem.
Production agentic AI demands infrastructure that can:
• Deploy agents reliably across environments
• Manage tool and API dependencies centrally
• Persist and recover state across long-running workflows
• Scale across cloud, on-prem, or hybrid without lock-in
This is exactly what Clarifai's Compute Orchestration was built for.
Deploy any model. On any compute. At any scale. With autoscaling, multi-environment support, and centralized control.
And when paired with the Clarifai Reasoning Engine, teams are seeing up to 2x performance at half the cost, because optimization does not stop at the model layer.
If you’re serious about taking agentic AI beyond prototypes, this is where the conversation shifts.
Explore how we’re building production-ready agentic AI:
www.clarifai.com/blog/clarifai-12.1-building-production-r… -
Clarifai
@clarifai
Long-running AI workflows produce more than predictions.
They produce state.
Model checkpoints. Training logs. Evaluation metrics. Preprocessed datasets. Configuration files.
With 12.1, we’re introducing Artifacts, versioned storage designed specifically for pipeline and workload outputs.
Each artifact:
• Supports immutable versions
• Stores large files efficiently in object storage
• Tracks metadata in the control plane for fast lookup
• Works seamlessly with Pipelines, CLI, and SDK
Why this matters:
Reproducibility. Save the exact weights and configs behind a result.
Checkpointing. Resume long-running jobs without recomputing.
Version control. Compare outputs across runs with precision.
As AI systems move beyond single requests into long-running, multi-step workflows, managing outputs becomes part of the infrastructure.
Artifacts make that explicit.
Learn more here: www.clarifai.com/blog/clarifai-12.1-building-production-r… -
Clarifai
@clarifai
Agentic AI is changing the cost and performance equation of building AI systems.
It’s no longer about a single model or a single prompt. Production systems today involve long-running workflows, agentic coordination, and sustained inference under real cost constraints.
In a recent conversation on Liftoff with Keith, Alfredo Ramos our CPTO, shared how Clarifai approaches this shift — from model optimization and token economics to hardware selection and orchestration at scale.
The discussion covers:
• Why open-weight models now support a significant share of real-world use cases
• How model optimization and infrastructure choices impact production performance
• Where orchestration becomes critical as systems grow more agentic
• What it takes to accelerate deployment while reducing cost
For teams moving from experimentation to production AI, these tradeoffs matter.
Watch the full conversation here: youtu.be/oDXiID_Ehxk -
Clarifai
@clarifai
Production agentic AI requires more than powerful models!
It requires infrastructure that can deploy agents reliably, manage the tools they depend on, and scale across any environment.
Clarifai 12.1 extends the core capabilities for agentic workloads:
✓ Deploy public MCP servers to give models tool-calling capabilities
✓ Store and version pipeline outputs with Artifacts
✓ Monitor and control long-running workflows through improved Pipeline UI
Available now in Public Preview.
Read the full release here: www.clarifai.com/blog/clarifai-12.1-building-production-r… -
Clarifai
@clarifai
GPU shortages are the headline.
But the deeper issue is structural.
AI workloads in 2026 look very different from two years ago. They’re longer-running. More agentic. More iterative. They don’t just spike once and finish. They stay active, coordinate tools, call models repeatedly, and operate under real cost pressure.
When workloads change, infrastructure stress shows up differently.
In many cases, the constraint isn’t absolute GPU supply. It’s fragmentation. Idle capacity in one place, saturation in another. Static allocation for dynamic systems.
That’s why the conversation is shifting from “How do we get more GPUs?” to “How do we use the ones we have more intelligently?”
Orchestration, scheduling, and workload design matter more than ever.
We wrote about where GPU pressure is actually coming from, and what teams should be thinking about as AI systems scale.
Read here: www.clarifai.com/blog/gpu-shortages-2026 -
Clarifai
@clarifai
Clarifai featured on Ramp’s Top Software Vendors for February 🚀
Ramp’s ranking is based on real spend and new customer adoption across thousands of companies.
At Clarifai, we’ve been focused on the foundations behind that adoption: compute orchestration that lets teams run models at scale across any infrastructure, pipelines for long-running, multi-step AI workflows, and the Clarifai Reasoning Engine built for agentic and reasoning workloads.
This is the infrastructure teams rely on when AI moves from experimentation into production.
More details in Ramp’s post: ramp.com/velocity/top-saas-vendors-on-ramp-february-2026 -
Clarifai
@clarifai
Why pipelines matter for real AI workloads
Once AI work stops fitting into a single request, things start to break.
Training runs, batch inference, evaluations, agent workflows, and data processing jobs don’t finish in seconds. They run asynchronously, span multiple steps, and often need hours or days to complete.
Trying to run this kind of work as one long execution makes it fragile. Failures are hard to recover from. Progress is hard to inspect. Retrying usually means starting over.
Pipelines exist to solve that exact problem.
Clarifai Pipelines let you define AI workflows as explicit, multi-step processes. Each step runs independently, can be monitored, retried, or resumed, and executes on managed compute inside the platform.
If your workload is long-running, multi-stage, or asynchronous, pipelines are the difference between something that works once and something you can actually operate in production.
👉 Learn how pipelines work on Clarifai: docs.clarifai.com/compute/pipelines/ -
Clarifai
@clarifai
Billing should not slow down AI work.
Clarifai uses a single Pay-As-You-Go credit system for self-serve usage. One balance across training, inference, workflows, and GPU-backed compute. No monthly minimums. No plan juggling.
Auto-Recharge keeps long runs uninterrupted. New users get a $5 credit to try real workloads.
More Details here: www.clarifai.com/blog/introducing-pay-as-you-go-credits -
Clarifai
@clarifai
You can now deploy public, open-source MCP servers directly on Clarifai ⚡
Upload an existing MCP server, deploy it once, and reuse it across models and workflows.
Agentic models can discover and call the same tools during inference, all from a shared, managed setup.
Clarifai already supports building and deploying custom MCP servers. This update makes it easier to bring in open-source ones without standing up your own service.
Here's an example of deploying a public browser MCP server:
github.com/Clarifai/runners-examples/tree/main/mcp/browse… -
Clarifai
@clarifai
🚀 Model Routing with Multi-Nodepool Deployments
You can now route a single model deployment across multiple nodepools.
With model routing, Clarifai lets you:
• Span deployments across multiple nodepools
• Handle traffic spillover automatically
• Scale across cloud, on-prem, or hybrid compute
• Improve availability without duplicating deployments
Routing logic is defined once in deployment_config.yaml. Clarifai handles the rest.
This turns deployments into resilient, multi-node services instead of single-cluster endpoints.
👉 Learn more here: docs.clarifai.com/compute/deployments/clusters-nodepools/ -
Clarifai
@clarifai
Clarifai 12.0 is live 🚀
Introducing Pipelines for long-running, multi-step AI workflows.
Run and monitor complex jobs directly on your compute without glue code.
Also in this release:
• Model routing across nodepools
• Expanded agentic MCP support
• Vercel AI SDK integration
• New Pay-As-You-Go plan
Read more here: www.clarifai.com/blog/clarifai-12.0-introducing-pipelines… -
Clarifai
@clarifai
H100 vs GH200 for AI workloads!
If you’re compute-bound and fit in GPU memory, H100 is usually the right call.
If memory limits or CPU↔GPU data movement are the bottleneck, GH200 starts to matter.
Full breakdown here:
www.clarifai.com/blog/nvidia-h100-vs.-gh200-choosing-the-… -
Clarifai
@clarifai
AI billing should not be unpredictable.
We heard the same issues from teams building on Clarifai: surprise charges, confusing plans, and jobs stopping mid-run.
That’s why self-serve billing is now Pay-As-You-Go with prepaid credits. No monthly commitments. Pay only for what you use.
Read more here: www.clarifai.com/blog/introducing-pay-as-you-go-credits -
Clarifai
@clarifai
We put together a breakdown of the Top 10 open-source reasoning models in 2026, focused on long context, multi-step reasoning, and agent execution.
Covers GPT-OSS-120B, GLM-4.7, Kimi K2 Thinking, Qwen3, Ministral 14B, and more.
Read it here:
clarifai.com/blog/top-10-open-source-reasoning-models-2026 -
Clarifai
@clarifai
Clarifai now uses a single Pay-As-You-Go credit system for self-serve usage.
One balance across training, inference, workflows, and GPU-backed Compute Orchestration. No monthly minimums. Fewer feature gates.
Every verified user gets a one-time $5 credit to try the platform.
Read more here: www.clarifai.com/blog/introducing-pay-as-you-go-credits -
Clarifai
@clarifai
Many image classification workflows require strict policies, edge-case handling, and consistent decisions at scale.
MM-Poly-8B is Clarifai’s multimodal model built for prompt-driven classification across images, text, audio, and video.
It’s well suited for moderation and safety-critical workflows where rules must be followed conservatively and consistently.
Define your policy in the prompt. Apply it reliably in production.
Try it out: clarifai.com/clarifai/main/models/mm-poly-8b -
Clarifai
@clarifai
Clarifai is now available as an inference provider in the Vercel AI SDK ⚡
You can use Clarifai-hosted models directly through the OpenAI-compatible interface in "ai-sdk/openai-compatible", with no changes to your application logic.
Once configured with a Clarifai Personal Access Token, you can call models like GPT-OSS-120B and other LLMs from the Clarifai Community using the same APIs you already use for text generation and streaming.
This makes it easy to build chat interfaces, agents, and real-time AI features while using Clarifai as the backend inference layer.
Full setup and examples below:
👉 ai-sdk.dev/providers/openai-compatible-providers/clarifai -
Clarifai
@clarifai
Reasoning workloads break most inference stacks. Long contexts, multi-step chains, and tool use drive up latency and cost fast.
The Clarifai Reasoning Engine adapts to how reasoning traffic behaves, optimizing kernels, batching, and memory in real time.
2x performance at half the cost for real-world agentic AI.
→ Learn how the Clarifai Reasoning Engine works: www.clarifai.com/products/model-inference -
Clarifai
@clarifai
Ministral-3-3B-Reasoning-2512 is now available on Clarifai! 🚀
The most efficient multimodal model in the Ministral 3 family:
✅ 4B total parameters for edge-optimized deployment.
✅ 256k Token Context for deep long-document analysis.
✅ Multimodal Input (Text + Image) supported.
✅ Reasoning post-trained for STEM/Agentic tasks.
Integrate this compact powerhouse today for low-resource agents!
Try out Ministral-3-3B-Reasoning-2512: clarifai.com/mistralai/completion/models/Ministral-3-3B-R… -
Clarifai
@clarifai
Teams rely on the Clarifai Control Center to understand how their models are performing, how workloads behave over time, and where their compute spend is going.
A recurring request has been the ability to take this data outside the dashboard for deeper analysis and internal reporting.
So we’ve added a simple but powerful update.
You can now export any chart in the Control Center as a CSV file. This makes it easy to pull usage metrics into your own dashboards, audit workflows, BI tools, or custom monitoring systems.
A small feature with a big impact for teams that need clearer visibility into their AI operations.
Try it out in the Control Center! -
Clarifai
@clarifai
Clarifai 11.11 is here 🚀
Trinity Mini is now available on Clarifai, a U.S. built open weight reasoning model trained on 10T tokens with strong scores on SimpleQA, MMLU and Math-500.
We also added Ministral-3-14B-Reasoning-2512 and GLM-4.6, along with Model Routing now in preview.
This release includes Control Center CSV export, platform updates and Python SDK improvements.
👉 Learn more in the release blog → www.clarifai.com/blog/clarifai-11.11-trinity-mini-a-new-u… -
Clarifai
@clarifai
Ministral-3-14B-Reasoning-2512 is now available on the Clarifai's Open AI Compatable API! 🚀
This ∼14B parameter, open-weight model is optimized for complex workloads:
✅ 256k Token Context for deep document analysis.
✅ Multimodal Input (Text + Image) supported.
✅ Reasoning Post-Trained for STEM/Coding tasks.
✅ Access via OpenAI-compatible API.
Start integrating the model for your agents today.
Try out Ministral-3-14B-Reasoning-2512: clarifai.com/mistralai/completion/models/Ministral-3-14B-… -
Clarifai
@clarifai
Ministral 3 14B Reasoning 2512 is now live on Clarifai. 🚀
It supports text and image inputs, handles 256k context, and is optimized for math, code, and complex reasoning.
Try it out in the Playground or access via Clarifai’s OpenAI-compatible API
👉 clarifai.com/mistralai/completion/models/Ministral-3-14B-… -
Clarifai
@clarifai
Need an efficient LLM for AI Agents?
Access Trinity Mini with an API 🚀
Trinity Mini (26B MoE / 3B Active) by @arcee_ai is now easily accessible on Clarifai.
✅ 128k Context for massive data analysis.
✅ Robust Function Calling for agents.
✅ Apache 2.0 open-weight license.
You can start testing its capabilities in the Clarifai Playground, or access it via the Clarifai OpenAI-compatible API to integrate it into your applications.
Try out Trinity mini on Clarifai: clarifai.com/arcee_ai/AFM/models/trinity-mini -
Clarifai
@clarifai
A lot of teams are realizing that their biggest agent bottlenecks come from the execution layer.
Slow token throughput, fragmented runtimes, and inconsistent environments make agents unreliable even when the models are strong.
At NeurIPS today, our CEO and Founder Matthew Zeiler walking through how a unified compute layer removes these performance cliffs and keeps agents fast, stable, and cost-efficient across cloud, on-prem, and edge setups.
If you’re at NeurIPS, join us:
📅 Dec 4, 2025
⏰ 4:00 PM
📍 San Diego Convention Center, Exhibit Hall A,B -
Clarifai
@clarifai
Inference is becoming the biggest cost in the AI stack.
Our Founder & CEO @mattzeiler joined @craigss on EyeOn AI Podcast to share why inference can’t be an afterthought and how unified compute orchestration cuts GPU spend for complex agents and long-context models.
Check it out👉 youtu.be/dAuWw7oixUI -
Clarifai
@clarifai
We’ve partnered with @arcee_ai to bring the Trinity family of powerful, end-to-end American-built open-weight models to the edge. 🚀
Trinity Mini and Nano are now available on Clarifai, proving that frontier models can be built with efficient capital and high performance.
- Trinity Nano (6B/1B active): Optimized for real-time, on-device and embedded use cases where efficiency is critical.
- Trinity Mini (26B/3B active): Tuned for multi-step agent workflows and function calling.
Why Trinity on Clarifai?
This launch is a win for developers seeking ownership and transparency.
By running on Clarifai, these models demonstrate our core value proposition: Twice the performance at half the price.
Key Differentiators:
• Competitive Benchmarks: Trinity Mini achieves 84.95% on MMLU (Zero Shot) and 92.10% on Math-500, demonstrating competitive capability in a compact MoE footprint (3B active parameters).
• Open Provenance: This is an end-to-end entirely American model, providing the legal certainty and control developers demand.
Ready to Build?
The Trinity family is available today with full Playground access and drop-in compatibility with OpenAI-spec clients and agent frameworks.
👉 Try the Trinity family on Clarifai: clarifai.com/arcee_ai/AFM/models/trinity-mini
👉 Read the press release: www.clarifai.com/press-release-clarifai-selected-inferenc… -
Clarifai
@clarifai
Clarifai will be at NeurIPS 2025 next week 🚀
We’re in San Diego to discuss the real bottlenecks in large-scale inference: token-heavy workloads, latency spikes, and low GPU utilization.
The Clarifai Reasoning Engine tackles these with adaptive performance and 2x throughput at half the cost.
Meet Clarifai at the booth T18
📆 Dec 2–5 | San Diego Convention Center
👉 www.clarifai.com/meet-clarifai-at-neurips-2025 -
Clarifai
@clarifai
Run SGLang models locally using Clarifai’s new SGLang Local Runners Toolkit 🚀
Initialize a model, run it on your machine, expose a local endpoint, and use the OpenAI-compatible API for inference.
👉 Checkout the SGLang Toolkit: docs.clarifai.com/compute/toolkits/sglang -
Clarifai
@clarifai
Twelve years ago, we started as one of the earliest AI platforms, long before AI platforms became a category. We built some of the first production AI APIs, video AI systems, model fine-tuning, vector databases, and on-prem deployment tools. Every step of the journey has been about giving developers and teams better ways to run AI at scale.
Today, that mission is more important than ever.
The next decade of AI will be defined by inference. Agentic systems are pushing infrastructure harder than anything we have seen. High token volumes, long context windows, multi-step reasoning, and tool use all require massive throughput and near-instant latency.
This is why we built the Clarifai Reasoning Engine—delivering 2x performance at half the cost for real world agentic AI.
It learns from the workloads running on it, adapting in real time to optimize kernels, batching, and memory use. For multi-step agent workflows, these optimizations compound and translate directly into:
⏱️ 65% faster time to first token.
⚡ 40% faster time to answer on production workloads.
📉 Better throughput and efficiency as patterns repeat.
Our goal is simple. Make large scale inference faster, more adaptive, and accessible for every team building the next generation of AI.
We are incredibly proud of the team and grateful for the customers and partners who have been with us on this journey.
Twelve years in, it still feels like day one. 🚀 -
Clarifai
@clarifai
Environment Secrets are now available on Clarifai 🔐
You can now store encrypted values directly in Clarifai and reference them as environment variables in your workflows. This makes it easier to manage API keys, credentials, and other sensitive configuration without hardcoding or exposing them in your codebase.
Environment Secrets provide a simple way to keep your workflows secure and maintainable, especially when working with multi-step agents or external integrations.
👉 Learn more here: docs.clarifai.com/control/authentication/environment-secr… -
Clarifai
@clarifai
Run GLM 4.6 via API!
GLM 4.6 brings unified reasoning, coding, and agentic capabilities in a single model, with a 200K context window, stronger coding performance on real-world tools, improved reasoning with tool use, more capable agent behavior, and refined writing quality.
You can now access GLM 4.6 through Clarifai’s OpenAI-compatible API.
👉 Try GLM 4.6 on Clarifai: clarifai.com/zai/completion/models/GLM_4_6 -
Clarifai
@clarifai
AI Infra 2025 by @IgniteGTM last week was incredible! 🚀
We spoke with teams running high-volume AI workloads in production, and many of the conversations pointed to the same bottlenecks: token-heavy reasoning steps, inconsistent latency during tool use, scaling across mixed clouds, and steadily rising inference costs.
A major highlight was the talk by Alfredo Ramos, our CPTO, on Scaling Agentic AI and Optimizing Tokenomics with Compute Orchestration and Local Runners.
In the session, Alfredo covered:
• Why agentic systems consume large volumes of tokens and how that impacts cost
• How Clarifai delivers faster inference at lower cost across clouds and hardware
• The role of multi-cloud and multi-architecture support
• How compute orchestration, GPU fractioning, and autoscaling improve efficiency
• How the Clarifai Reasoning Engine adapts to workloads to improve latency and performance
If you want a clear look at how large-scale agents run in production and how to optimize their cost model, this talk is worth watching.
👉 Watch the full session here: youtu.be/9krz_jAA9JQ -
Clarifai
@clarifai
Join us for a research seminar with M. Riley Knoedler on using UAS and AI to detect bats in thermal video at wind energy sites. 🚀
A custom UAS and CV pipeline processed 4,000+ videos and identified segments with potential bat activity, reducing review effort and improving detection consistency.
🗓 Nov 21, 2 PM ET
👉 Register here: www.clarifai.com/research-seminar-a-custom-uncrewed-aircr… -
Clarifai
@clarifai
Structured outputs with Clarifai’s OpenAI-compatible API!
Generate schema-validated JSON from models like GPT-OSS-120B and plug the results directly into your applications.
Define a schema, send it in the request, and get consistent structured output back.
👉 -
Clarifai
@clarifai
Clarifai 11.10 is here! 🚀
In this release, we are excited to introduce Single-Click Deployment, which removes all manual cluster setup and lets you deploy models instantly.
Key Highlights:
• Single-Click Deployment: Deploy models instantly without any manual cluster setup. Clarifai recommends the right instance type and automatically creates clusters when needed.
• New Models: DeepSeek-OCR for high-precision document processing and GLM-4.6 for unified reasoning, coding, and agentic intelligence.
• Control Center: Unified Ops and token reporting across all models for consistent visibility.
• Structured Outputs: Produce schema-validated JSON from any OpenAI-compatible model using Pydantic schemas.
👉 Explore the full release updates here: www.clarifai.com/blog/clarifai-11.10-seamless-model-deplo… -
Clarifai
@clarifai
New Cookbook — Conditional Routing Agent 🤖
Different tasks need different models. A single LLM isn’t always the best fit for every query.
The Conditional Routing Agent Workflow shows how to build a system that automatically selects the right model for each task using Clarifai’s OpenAI-compatible API.
It routes inputs based on model strengths:
• Coding → Qwen3-Coder-30B
• Reasoning → GPT-OSS-120B
• Conversation/Summarization → Llama-3.2-3B-Instruct
A lightweight Router LLM evaluates the input, outputs structured JSON specifying the model to use, and runs the workflow end-to-end.
This notebook demonstrates how to implement clean, modular multi-model orchestration with Clarifai.
👉 Checkout the Cookbook here: github.com/Clarifai/examples/blob/main/agents/conditional… -
Clarifai
@clarifai
Reasoning Engine that adapts to your workloads 🤖
Inference workloads are rarely consistent. Latency, token mix, and batch density fluctuate as agents reason, plan, and call tools.
The Clarifai Reasoning Engine adapts to these workload shifts in real time. It optimizes kernels, batching, and memory utilization dynamically, improving throughput and efficiency with every run.
For repetitive or structured agentic tasks, this adaptive behavior compounds, delivering faster inference and lower cost without sacrificing accuracy.
👉 Learn more: www.clarifai.com/products/model-inference -
Clarifai
@clarifai
Clarifai at SuperCompute 2025 🚀
Reasoning workloads demand extreme speed and efficient scaling. Most GPU systems struggle to keep up.
The Clarifai Reasoning Engine delivers 2x throughput at half the cost and optimizes kernels, batching, and memory in real time.
👉 Find time with the team here: www.clarifai.com/super-compute-2025 -
Clarifai
@clarifai
GLM-4.6 is now live on Clarifai! 🚀
It brings a 200K context window, stronger reasoning and coding abilities, and improved agent behavior with tool-use support.
Outperforms GLM-4.5 across 8 benchmarks.
Try it in the Playground or access via API → clarifai.com/zai/completion/models/GLM_4_6 -
Clarifai
@clarifai
Agentic workloads need low latency and efficient scaling!
In benchmarks by @ArtificialAnlys, Clarifai’s Reasoning Engine running GPT-OSS-120B achieved 0.36s TTFT, 500+ tokens per second throughput, and $0.16 per million tokens on standard GPUs
If you’re building systems that need real-time reasoning or rapid multi-turn responses, this is built for you.
👉 Try GPT-OSS-120B here: clarifai.com/openai/chat-completion/models/gpt-oss-120b -
Clarifai
@clarifai
Most teams spend more time managing GPUs than building models.
Run large models like GPT-OSS-120B or your own on Clarifai’s Reasoning Engine, optimized for inference, scaling, and utilization across NVIDIA B200s, GH200s, and H100s.
Reach out here → www.clarifai.com/compute-orchestration/gpu-hosting -
Clarifai
@clarifai
Run LM Studio models locally with a public API! 🔥
@lmstudio makes it easy to run and experiment with open-source models on your local machine. With Clarifai Local Runners, you can now securely connect those models to your apps or workflows.
You can serve any LM Studio model directly from your laptop and get a Public API endpoint instantly. No hosting, tunneling, or extra setup needed.
Everything runs on your local machine while Clarifai handles routing, authentication, and integration.
In this blog, learn how to:
• Initialize and configure LM Studio models with the Clarifai CLI
• Run Local Runners and connect them to your LM Studio runtime
• Send inference requests using the OpenAI-compatible API or the Clarifai SDK
Read the full guide here: www.clarifai.com/blog/run-lm-studio-models-locally -
Clarifai
@clarifai
Clarifai at AI Infra Live 2025!
The Clarifai team will be in San Francisco this week to showcase how the Clarifai Reasoning Engine powers high-performance AI workloads at scale.
From efficient GPU inference to large-scale deployment, the Reasoning Engine helps teams run production-grade AI with speed and precision.
If you are attending AI Infra, meet the team and explore how Clarifai can help streamline your AI infrastructure.
📅 November 7, 2025 | San Francisco
👉 Book a meeting: www.clarifai.com/ai-infra-live-event-2025 -
Clarifai
@clarifai
🚀 New E-book for AI Leaders
As enterprises move from basic generative AI to agentic systems, the complexity of managing compute, costs, and innovation grows fast.
That’s why we created an e-book to help you navigate these challenges, from cost optimization to building scalable and efficient AI infrastructure.
Learn how to design, deploy, and manage modern AI systems built for speed and resilience.
📘 Download the free e-book here: www.clarifai.com/ebook-clarifai-executive-guide-2025 -
Clarifai
@clarifai
Run DeepSeek-OCR with an API⚡
Extract structured text, tables, and formulas from complex docs with high accuracy.
Built on SAM + CLIP + MoE for speed and precision.
Read the full guide:
www.clarifai.com/blog/run-deepseek-ocr-with-an-api -
Clarifai
@clarifai
Run vLLM Models Locally with a Secure Public API!
vLLM is a high-throughput inference engine optimized for large language models.
With Clarifai Local Runners, you can serve vLLM models directly from your own machine while exposing them through a secure public API.
Run models privately, test them fast, and keep full control over your environment.
Read the full guide → www.clarifai.com/blog/run-vllm-models-locally-with-a-secu… -
Clarifai
@clarifai
🚀 We’re live at NVIDIA GTC Washington, D.C. with @Vultr, showcasing what next-gen AI inference really looks like.
The Clarifai Reasoning Engine is pushing new limits on Vultr’s GPU clusters, delivering 544 tokens/sec, 0.36s to first token, and an unmatched $0.16 per million tokens on GPT-OSS-120B, as benchmarked by Artificial Analysis.
That’s the performance you can feel — real speed, and efficiency for Agentic AI.
If you’re at NVIDIA GTC Washington, D.C., stop by the Vultr booth #453 to see it in action.
👉 Learn more in the press release: www.clarifai.com/press-release-clarifai-and-vultr-showcas… -
Clarifai
@clarifai
DeepSeek OCR is now live on the Clarifai! 🚀
- 96%+ OCR precision at 9–10× compression
- Processes 200K+ pages/day on a single A100 GPU
- Efficient DeepEncoder architecture for large-scale OCR
👉 Try the model here: clarifai.com/models/deepseek-ocr -
Clarifai
@clarifai
Clarifai at AI Infra live 2025 by @IgniteGTM! 🚀
We will be at the event to showcase how the Clarifai Reasoning Engine delivers fast, accurate GPU inference for large-scale AI workloads.
📅 Nov 7 | San Francisco
👉 www.clarifai.com/ai-infra-live-event-2025 -
Clarifai
@clarifai
Clarifai at NVIDIA GTC Washington, D.C! 🚀
We recently launched the Clarifai Reasoning Engine for Agentic AI inference, delivering cost and efficiency gains in real production workloads.
Benchmarks from Artificial Analysis show it’s the fastest GPU-based inference engine, outperforming even top ASICs.
Join the Clarifai team to see it in action.
📅 Oct 27–29, 2025 | Washington, DC
Book a meeting: www.clarifai.com/nvidia-gtc-dc-2025 -
Clarifai
@clarifai
Run @huggingface models locally and serve them through a Public API! 🤗
No uploads, no cloud setup. Just secure local inference with a public API.
Clarifai’s Local Runners make this possible. Think of it as ngrok for AI models, letting you serve models, MCP servers, or agents directly from your own hardware while Clarifai handles routing, security, and integration.
Here’s what you can do with Local Runners:
✅ Run models on your own machines
✅ Keep data and compute private, nothing leaves your machine
✅ Skip API setup, Clarifai provides the endpoint and routing
✅ Prototype and iterate quickly in real-world pipelines
Read the full guide to learn how to run Hugging Face models locally with the new Clarifai Toolkit 👇
🔗 www.clarifai.com/blog/run-hugging-face-models-locally -
Clarifai
@clarifai
Access GPT-OSS 120B directly through Clarifai’s OpenAI-compatible API and build agents using the OpenAI Agents SDK.
You can easily plug Clarifai models into your existing workflow and start creating reasoning agents that use tools and respond intelligently.
The Clarifai Reasoning Engine powers fast and efficient execution of GPT-OSS 120B. It delivers high token throughput and low time-to-first-token.
Benchmarks from Artificial Analysis on GPT-OSS 120B show that the Clarifai Reasoning Engine delivers the fastest GPU-based inference measured.
Try out GPT-OSS 120B: clarifai.com/openai/chat-completion/models/gpt-oss-120b
Read more here: docs.clarifai.com/compute/agents/build -
Clarifai reposted
Clarifai
@clarifai
We’re at #pytorch2025 in San Francisco! 🔥
Meet the Clarifai team at booth B15 (Oct 22–23) with @mattzeiler (CEO & Founder) and the rest of the team.
Let’s talk benchmarks, agents, and inference.
👉 clarifai.com/pytorch-conference-2025 -
Clarifai
@clarifai
We’re at #pytorch2025 in San Francisco! 🔥
Meet the Clarifai team at booth B15 (Oct 22–23) with @mattzeiler (CEO & Founder) and the rest of the team.
Let’s talk benchmarks, agents, and inference.
👉 clarifai.com/pytorch-conference-2025 -
Clarifai
@clarifai
🤗 We’ve added a Hugging Face Toolkit to the Clarifai CLI.
This makes it simple to initialize, customize, and serve Hugging Face models through Local Runners.
You can now download and run supported Hugging Face models directly on your own hardware, such as laptops, workstations, or edge boxes, while exposing them securely via Clarifai’s public API.
Your model runs locally, your data stays private, and Clarifai handles routing, authentication, and governance.
Why this matters:
⚙️ Use local compute – Run open-weight models on your own GPUs or CPUs while keeping them API-accessible.
🔒 Preserve privacy – All inference happens locally; only metadata flows through Clarifai’s secure control plane.
⚡ Skip manual setup – Initialize a ready-to-run model directory with one CLI command.
All it takes is a single command:
𝐜𝐥𝐚𝐫𝐢𝐟𝐚𝐢 𝐦𝐨𝐝𝐞𝐥 𝐢𝐧𝐢𝐭 --𝐭𝐨𝐨𝐥𝐤𝐢𝐭 𝐡𝐮𝐠𝐠𝐢𝐧𝐠𝐟𝐚𝐜𝐞 --𝐦𝐨𝐝𝐞𝐥-𝐧𝐚𝐦𝐞 𝐐𝐰𝐞𝐧/𝐐𝐰𝐞𝐧2-0.5𝐁
Read more here: docs.clarifai.com/compute/local-runners/hf/ -
Clarifai
@clarifai
Clarifai 11.9 is here! 🚀
Here are the key highlights:
• Clarifai Reasoning Engine – faster, more efficient agentic AI inference
• Run models locally – Hugging Face, vLLM & LM Studio toolkits via Local Runners
• New Qwen reasoning models – optimized for advanced reasoning workloads
• Expanded cloud options – B200 & GH200 instances for scalable deployments
Learn more about all the updates in the release blog here: www.clarifai.com/blog/clarifai-11.9-introducing-clarifai-… -
Clarifai
@clarifai
Heading to AI Infra Live 2025!
See how the Clarifai Reasoning Engine powers agentic AI workloads with GPU performance verified by Artificial Analysis.
Looking to scale AI workloads cost-efficiently? Connect with the team.
📆 Nov 7, 2025
👉 clarifai.com/ai-infra-live-event-2025 -
Clarifai
@clarifai
How do top inference providers perform on GPT-OSS-120B?
We compare throughput, latency, and cost across platforms using benchmarks from Artificial Analysis.
Discover which providers excel at reasoning-heavy workloads.
Full comparison here: www.clarifai.com/blog/top-llm-inference-providers-compared -
Clarifai
@clarifai
Enterprise-Grade GPU Hosting for AI Models is here!
Run GPT-OSS-120B and custom models on NVIDIA B200s, GH200s, and H100s — optimized for large-scale inference, reasoning, and AI agents.
Explore the full lineup and request GPUs for your next workload:
www.clarifai.com/compute-orchestration/gpu-hosting -
Clarifai
@clarifai
Heading to #pytorchconf2025? So are we!
Meet Clarifai CEO @mattzeiler and Developer Advisor Rob Spectre at booth B15 in San Francisco.
We’ll be showcasing the Clarifai Reasoning Engine, benchmarked by Artificial Analysis for record GPU inference speed.
👉 Let’s talk performance and agentic AI: clarifai.com/pytorch-conference-2025 -
Clarifai
@clarifai
Access GPT-OSS 120B with an API!
Benchmarks by @ArtificialAnlys confirm the GPT-OSS 120B model responsiveness, throughput, and efficiency on the Clarifai Reasoning Engine, making it production-ready for real-world agentic workflows.
Try it out here: clarifai.com/openai/chat-completion/models -
Clarifai
@clarifai
🧠 Qwen3-Next-80B-A3B-Thinking on Clarifai
80B parameters, with only 3B active per step — built for deep reasoning at breakthrough efficiency.
Try it out here: clarifai.com/qwen/qwen3/models/qwen3-next-80B-A3B-Thinking -
Clarifai
@clarifai
PyTorch Conference 2025 is almost here.
Clarifai will be at Booth #b15 on October 22–23 in San Francisco.
We’ll be showcasing the Clarifai Reasoning Engine, built for high-throughput, long-running reasoning workloads.
If you’d like dedicated time, schedule a meeting with CEO @mattzeiler or Developer Advisor Rob Spectre here:
👉 www.clarifai.com/pytorch-conference-2025
#pytorchconference -
Clarifai
@clarifai
When Agents moves into production, speed matters as much as accuracy.
Agentic AI relies on real-time reasoning. Even small delays can break workflows and hurt adoption.
In tests by @ArtificialAnlys, the Clarifai Reasoning Engine achieved a 3.6s time to first answer on GPT-OSS 120B, showing the responsiveness needed for multi-step agent workflows.
👉 Try it on Clarifai: clarifai.com/openai/chat-completion/models/gpt-oss-120b -
Clarifai
@clarifai
Agentic workloads burn through tokens fast. They need ultra-low latency and scalable economics.
In benchmarks by @ArtificialAnlys on GPT-OSS 120B, the Clarifai Reasoning Engine achieved 0.36s TTFT, 500+ tokens per second throughput, and $0.16 per million tokens (blended), all on standard GPUs.
Built for agentic AI that demands speed, scale, and reliability.
👉 Try out GPT-OSS-120B clarifai.com/openai/chat-completion/models/gpt-oss-120b -
Clarifai
@clarifai
Clarify Research Seminar 🎙️
Why do overparameterized neural nets still generalize so well? Is it the bias of SGD or the architecture itself?
Join Amit Peleg (PhD, Univ. of Tübingen) as he breaks it down.
📅 Oct 3 | 12 PM ET | 9 AM PT
Register here👉 www.clarifai.com/research-seminar-bias-of-stochastic-grad… -
Clarifai
@clarifai
Clarifai is heading to PyTorch Conference 2025!
We’ll be at booth B15.
If you’re tired of duct-taping AI tools together, come see how Clarifai runs the whole lifecycle in one place with faster compute, lower costs, and hardware flexibility.
@mattzeiler (CEO & Founder) and Rob Spectre (Dev Advisor) will both be there.
Grab time here: www.clarifai.com/pytorch-conference-2025 -
Clarifai
@clarifai
Throughput is the real unlock for Agentic AI!
When reasoning models burn through tokens, speed decides whether they’re practical.
Clarifai Reasoning Engine, tested on GPT-OSS 120B by Artificial Analysis, delivered 544 tokens/sec on standard GPUs, the fastest GPU-based inference ever recorded.
For the first time, GPUs match (and even beat) specialty non-GPU accelerators — proving that high-throughput reasoning doesn’t have to mean hardware lock-in.
Try out GPT-OSS-120B model here: clarifai.com/openai/chat-completion/models/gpt-oss-120b -
Clarifai
@clarifai
Thrilled to see @TechCrunch cover the launch of the Clarifai Reasoning Engine.
Benchmarked on GPT-OSS 120B, we hit:
- 544 tokens/sec throughput
- 3.6s time-to-first-answer
- $0.16 per million tokens
Reasoning workloads burn tokens at extreme speeds. With Clarifai, developers now get record throughput, ultra-low latency, and unmatched cost efficiency — all on standard GPUs.
Read more in TechCrunch: techcrunch.com/2025/09/25/clarifais-new-reasoning-engine-… -
Clarifai
@clarifai
🚀 Clarifai at Comcast Connect 2025!
Our Founder & CEO, @mattzeiler, is heading to Comcast Connect 2025 to explore the cutting edge of AI and Connectivity.
We’ll be diving into the future of telecom, including:
- Intelligent Network Infrastructure
- Predictive Operations
- Next-Gen Digital Experiences
Also on the ground: Alfredo Ramos (CPTO) and Rachael Kloek (Sr. AE)—ready to connect with attendees. See you there!
🔗 www.clarifai.com/comcast-connect -
Clarifai
@clarifai
🎉 Big news: Qwen3-Next-80B-A3B-Thinking is live on the Clarifai Platform!
This reasoning-optimized LLM delivers flagship-level problem solving with just 3B active parameters per step—efficient, fast, and cost-effective.
Key Highlights:
• Sparse Mixture-of-Experts → lower FLOPs/token
• Hybrid Attention → stable long-context handling
• Benchmarks: beats Gemini-2.5-Flash-Thinking & Qwen3-32B
Try it now in the Clarifai Playground or via our OpenAI-compatible API:
🔗 clarifai.com/qwen/qwen3/models/qwen3-next-80B-A3B-Thinking -
Clarifai
@clarifai
📣 ANNOUNCEMENT DAY AT CLARIFAI 📣
Today, we're thrilled to introduce the Clarifai Reasoning Engine for Agentic AI inference – purpose-built for reasoning workloads and optimized for performance at scale.
In independent benchmarks by @ArtificialAnlys on GPT-OSS 120B, Clarifai Reasoning Engine delivered record-setting results on standard GPUs:
- 544 tokens/sec throughput → the fastest GPU-based inference
- 0.3s time-to-first-token → near-instant responsiveness
- $0.16 per million tokens (blended) → the lowest cost in the industry
Clarifai is the first and only provider to combine top-tier speed with unmatched cost efficiency for large-scale reasoning and agentic workloads.
And it’s not just GPT-OSS 120B. Bring your own reasoning models → Clarifai will run them with the same unparalleled speed and economics.
👉 Read all about it here: www.clarifai.com/press-release-clarifai-launches-reasonin… -
Clarifai
@clarifai
Clarifai is excited to join Comcast Connect 2025!
Our CEO @mattzeiler will speak on the AI + Connectivity panel, joined by CPTO Alfredo Ramos and Sr AE Rachael Kloek.
Let’s connect at the event.
Book a slot with the team here: www.clarifai.com/comcast-connect -
Clarifai
@clarifai
Smarter agents and faster workflows with DeepSeek V3.1 on Clarifai ⚡
The most advanced open-weight model from DeepSeek, built for long-context tasks, tool use, and enterprise-ready performance.
Access via the API or try out the model in Playground:
👉 clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-V3_1 -
Clarifai
@clarifai
Building AI agents isn’t just about performance. It’s also about security + isolation!
At @daytonaio AI Builders Meetup, we showed how to safely run complex workloads with Clarifai's Compute Orchestration + Daytona sandboxes.
Check it out here: 👉 youtu.be/S9pMcLzqbgo -
Clarifai reposted
Daytona
@daytonaio
🎤Live from Daytona AI Builders Meetup @github Office in San Francisco last week.
Patrick Lundquist showed how @clarifai orchestration with @daytonaio sandboxes enable safe, isolated execution of AI models.
🎥 Link to the talk in comment ⬇️ -
Clarifai
@clarifai
Looking for a scalable, flexible AI infrastructure partner?
Clarifai was named a Challenger and Fast Mover in the GigaOm Radar for AI Infrastructure v1 Report recognized for:
- Transparent pricing
- Rich pre-built models
- Compute orchestration that maximizes efficiency across workloads
Read the full report here: www.clarifai.com/report/gigaom -
Clarifai
@clarifai
We’ve made billing more predictable and transparent!
• Developer & Essential plans → $100 monthly spending limit
• Professional plan → $500 monthly spending limit
Need higher limits? Reachout to the team.
We’ve also introduced a simple credit card pre-authorization step ($50–$500 depending on plan). It’s fully refunded within 7 days and ensures a smoother, more secure verification process.
These changes give you more control and clarity over your usage from day one
👉 See full details on our Pricing page: www.clarifai.com/pricing -
Clarifai
@clarifai
Grok Code Fast 1 has been making waves since launch 🚀
⚡ 256k context, 92–190 tok/sec
🧠 314B MoE architecture
🤖 Agentic workflows + multi-step automation
📊 70.8% SWE-Bench
Run it on Clarifai for just $0.25/M in & $1.875/M out.
👉 clarifai.com/xai/playground?model=grok-code-fast-1__15fb1… -
Clarifai
@clarifai
. @ArtificialAnlys independently evaluated providers serving GPT-OSS-120B across latency, throughput, and cost. ⚡
Clarifai’s Compute Orchestration delivered 0.27s time-to-first-token (TTFT), 313 tokens per second throughput, and a blended cost of just $0.16 per million tokens.
These results place Clarifai in the benchmark’s “most attractive quadrant” for speed and cost-efficiency — making it easier for engineering teams to scale GPT-OSS-120B in production.
👉 Read the full breakdown here: www.clarifai.com/blog/artificial-analysis-benchmarks-on-g… -
Clarifai
@clarifai
Ready to build the future of AI?
Meet Clarifai's CEO @mattzeiler and Developer Advisor Rob Spectre at #pytorchconf2025!
We'll be at booth B15 in San Francisco. Come join and talk with us on Compute and Inference!
www.clarifai.com/pytorch-conference-2025 -
Clarifai
@clarifai
New in Control Center 🚀
Control Center is a unified interface for centralized visibility into your platform usage.
In the latest release we’ve made Control Center more flexible:
📊 Resize charts to half-size for smooth side-by-side comparisons.
💡 Smarter charts: average costs displayed correctly, weekly aggregation on longer ranges, and clear messages when data isn’t available.
🔗 Cross-links between compute cost and usage charts for a complete view of your AI infrastructure.
These updates make it easier to analyze, compare, and manage your workloads at a glance.
Reaad more about Control Center here: docs.clarifai.com/control/control-center/ -
Clarifai
@clarifai
Clarifai has always focused on building infrastructure that lasts, and the latest benchmarks show why that matters. ⚡
Artificial Analysis tested GPT-OSS 120B on our platform and validated a rare combination: speed, efficiency, and flexibility in a single system.
Faster time to first token, strong throughput, and optimized GPU utilization aren’t just numbers — they’re what make reasoning agents, coding assistants, and research pipelines work at scale.
At $0.16 per million tokens (blended), our pricing undercuts the $0.26–$0.28 cluster while maintaining top-tier performance.
🔗 Read more about the benchmarks here: www.clarifai.com/press-release-artificial-analysis-ranks-…
👉 Try out GPT-OSS-120B here: clarifai.com/openai/chat-completion/models/gpt-oss-120b -
Clarifai
@clarifai
Clarifai 11.8 is here! 🔥
Clarifai Compute Orchestration just ranked top tier in Artificial Analysis benchmarks with GPT-OSS-120B, showing strong results in speed, scale, and cost efficiency.
Along with this, we’ve added important updates across the platform:
• Local Runners let you run models like GPT-OSS-20B with Ollama and serve them through a public API while keeping data local
• Control Center now has resizable and smarter charts for a clearer view of cost and usage
• Pricing is more transparent with spending limits and simple pre-authorization
Read the full release blog here: www.clarifai.com/blog/clarifai-11.8-gpt-oss-120b-benchmar… -
Clarifai
@clarifai
We’re excited to share that @ArtificialAnlys has validated Clarifai's Compute Orchestration with GPT‑OSS‑120B as one of the top‑performing GPU inference platforms globally. 🚀
What the data shows:
• 0.27s Time‑to‑First‑Token (TTFT) → tokens start streaming almost instantly.
• 313 output tokens/sec → among the very fastest measured.
• $0.16 per 1M tokens (blended) → this undercuts the $0.26–$0.28 cluster while maintaining top‑tier performance.
Placed in the benchmark’s “most attractive quadrant” for high speed + low price.
This is the payoff from years of building and operating inference infrastructure:
👉 Press Release here: www.clarifai.com/press-release-artificial-analysis-ranks-…
Clarifai
@clarifai
Full benchmarks here:
artificialanalysis.ai/models/gpt-oss-120b/providers?endpo… -
Clarifai
@clarifai
Deploying your first AI model doesn’t need to be complicated!
Most people think deploying an AI model means wrestling with DevOps and GPU configs!
With Clarifai’s Compute Orchestration, it works differently. You can spin up dedicated infrastructure and start running predictions in minutes.
Here’s how to deploy your own model:
1. Create an account to access the platform’s powerful AI deployment capabilities
2. Pick a model available on the community such as GPT-OSS-120B or Gemma-3 or bring your own model
3. Select your prefered compute from options like L40s, A100s, or H100s depending on your workload
4. Deploy and let Clarifai provision the infrastructure automatically
5. Run inferences in the Playground and test the model with real prompts
Instead of managing clusters and configs, you focus directly on the model and its outputs.
Learn how to deploy your first model step by step → docs.clarifai.com/getting-started/first-deployment -
Clarifai
@clarifai
AI is only as powerful as the infrastructure it runs on!
But here’s the reality: most enterprises are still overspending on workloads, stuck with bottlenecks, and struggling to scale innovation.
That’s why we created the Executive Guide to Optimizing AI Workloads — a practical playbook for leaders who want to cut costs, streamline operations, and stay ahead in the AI race.
👉 Get the guide here: www.clarifai.com/ebook-clarifai-executive-guide-2025 -
Clarifai
@clarifai
Testing models one by one is slow!
Now you can compare multiple models side by side in the Playground.
Run the same prompt across models, see differences in speed, quality, and style in one view.
👉 Try it now: clarifai.com/playgroundThis media couldn't be displayed. See it on the original post. -
Clarifai
@clarifai
Run advanced open source models like GPT-OSS-20B, Gemma-3-270M, and more directly on your hardware with Clarifai Local Runners.
Now with Ollama support, so you can run any Ollama model locally and access it via public API.
Developer Plan → $1/month
🎯 www.clarifai.com/pricing -
Clarifai
@clarifai
Mastering Reasoning Effort in the Playground!
When working with reasoning models, the Reasoning Effort parameter can completely change how the model behaves.
• Low Effort → Fast, lightweight, fewer tokens. Perfect for prototyping.
• Medium Effort → Balanced speed + depth. Great for experimentation.
• High Effort → Slower, but richer and more accurate answers.
By adjusting this setting, you can decide whether you want lightweight responses for quick iteration or deeper reasoning when accuracy and nuance matter most.
👉 Try it out yourself in the playground: clarifai.com/playground -
Clarifai
@clarifai
Not all LLM inference frameworks are built the same!
We compared SGLang, vLLM, and TensorRT-LLM on GPT-OSS-120B (2× H100 GPUs).
Here’s how they stack up on latency, throughput, and scaling.
Read the full benchmark breakdown:
www.clarifai.com/blog/comparing-sglang-vllm-and-tensorrt-… -
Clarifai
@clarifai
Missed our research seminar on Mapping biodiversity at very-high resolution in Europe?
The recording is now available.
Lukáš Picek (Inria) presented GeoPl@ntNet, an open platform mapping plant biodiversity at 50×50m resolution with multimodal deep learning.
Watch here → -
Clarifai
@clarifai
We’re excited for Citi’s Global TMT Conference!
Our CEO @mattzeiler joins the "Enterprise Applications in AI" panel on Sep 4th at 8:10–8:45 AM EDT
Hear how Clarifai is shaping the future of enterprise AI with agentic AI + compute orchestration.
If you’re in NYC, don’t miss it! -
Clarifai
@clarifai
AI at scale requires more than strong models, it requires the right infrastructure.
The GigaOm Radar for AI Infrastructure highlights the top platforms for training, deployment, and orchestration.
Clarifai is featured among the leaders.
Read the full report here: → -
Clarifai reposted
Turing Post (Ksenia Se)
@TheTuringPost
If you need to bridge your local models to the cloud – try @clarifai’s Local Runners
They allow you to:
• Run models on your laptop, servers or VPC clusters
• Build complex pipelines with other models, agents and tools
• Test changes instantly
• Debug faster
And all of this securely.
Connect models with just one command.
Start for free → www.clarifai.com/products/local-runners -
Clarifai reposted
Vultr
@Vultr
Enterprises are investing heavily in AI, but scaling and deploying models is still a major hurdle.
@clarifai CEO Matt Zeiler shares how full-stack AI orchestration with Vultr cuts complexity and slashes compute costs by up to 90%. 🧠💡 #ai #cloudcomputing
youtu.be/LMmC8PPZBhk -
Clarifai
@clarifai
Our CEO Matt Zeiler joins the Enterprise Applications in AI panel at Citi’s Global TMT Conference (NYC).
Hear how Clarifai helps enterprises scale agentic AI into real impact.
📅 Sept 4 | 8:10–8:45 AM EDT
clarifai.com/citis-global-tmt-conference-2025 -
Clarifai
@clarifai
The shift from generative AI to Agentic AI demands more than powerful models, it requires resilient infrastructure!
Yet only 13% of CIOs say their orgs are ready for tech disruptions.
Our new Executive Guide shows how to optimize AI workloads, cut hidden costs and accelerate innovation.
👉 Download it here: www.clarifai.com/ebook-clarifai-executive-guide-2025 -
Clarifai
@clarifai
How does GPT-OSS stack up against GLM-4.5, Qwen3, DeepSeek, and Kimi K2?
Benchmarks like MMLU-Pro, AIME, SWE-bench, GPQA, TAU-bench, and BFCL-v3 highlight the trade-offs:
- GPT-OSS = efficient reasoning
- GLM-4.5 = best at agentic workflows
- DeepSeek / Qwen3 = hybrid reasoning at scale
- Kimi K2 = compact coding focus
👉 Read more: www.clarifai.com/blog/openai-gpt-oss-benchmarks-how-it-co… -
Clarifai
@clarifai
Using OpenAI Responses API with Clarifai!
The OpenAI-compatible Responses API is available on Clarifai.
You can use the standard OpenAI client with Clarifai-hosted models and get access to:
✅ Text and multimodal support
✅ Streaming responses
✅ Tool calling
✅ Image generation
Checkout the Documentation here:
docs.clarifai.com/compute/inference/open-ai/ -
Clarifai
@clarifai
Clarifai Research Seminar this Week!
Join us for a session with Lukáš Picek (Inria, France and incoming Fulbright Scholar at MIT) as he presents GeoPl@ntNet, an open-access platform mapping European plant biodiversity at 50×50 meter resolution.
📅 August 29 | 12 PM ET | 9 AM PT
Discover how multimodal AI and deep learning are transforming biodiversity monitoring, research, and conservation.
👉 Register here: www.clarifai.com/research-seminar-mapping-biodiversity-at… -
Clarifai
@clarifai
Excited to share that our CEO, @mattzeiler, will be at Citi’s Global TMT Conference in NYC (Sep 3 - 5)!
Learn how Clarifai is helping enterprises bring agentic AI to life.
If you’re there, let’s connect: clarifai.com/citis-global-tmt-conference-2025 -
Clarifai
@clarifai
Missed our live webinar with @RelevanceAI_ ? 🚀
Watch the recording to learn how to build enterprise-grade AI agents that act, plan & execute securely at scale.
Discover best practices, deployment-ready architectures & real-world use cases
👉 www.clarifai.com/webinar-enterprise-grade-agentic-ai -
Clarifai
@clarifai
Agentic AI is pushing enterprises to rethink how we evaluate intelligence!
It’s not just about benchmark scores, it’s about how systems actually perform and feel in real-world use.
As Matt Zeiler, our CEO, explains: two models may score the same, but they won’t always deliver the same experience. That human “feeling” isn’t captured in today’s metrics, yet it’s critical to trust and adoption.
The next phase of AI progress will be defined by better ways to measure alignment, usefulness, and impact beyond the numbers. Looking forward to seeing where the industry is headed.
More thoughts on this shift here: fastcompany.com/91384747/databricks-wants-enterprises-to-…
Matt Zeiler
@mattzeiler
Agentic AI is reshaping how we measure intelligence!
It’s not just scores, but how models perform and feel in real use.
Two models can score the same yet deliver different experiences.
The future lies in measuring alignment, usefulness, and impact. -
Clarifai
@clarifai
Choosing between NVIDIA H100 vs B200 for your next AI deployment?
In our latest blog, we compare:
• The architectural evolution from Hopper to Blackwell
• Core specification and design differences
• GPT-OSS-120B inference performance across both GPUs
Read the blog here: -
Clarifai
@clarifai
🚀 Clarifai 11.7 is here!
The release highlights the GPT-OSS benchmarks on NVIDIA B200, and H100 GPUs using vLLM, SGLang, and TensorRT-LLM
Our research team tested GPT-OSS across single-request and high-concurrency workloads—check out the full results in the release blog below.
Other Key highlights:
• Developer Plan: Run up to 5 Local Runners with unlimited hours for just $1/month
• New models: GPT-OSS-120b, GPT-5 Mini & Nano, Qwen3-Coder-30B
• Ollama support: Download, run, and expose models locally via API
• Playground improvements: Compare models side by side, enhanced controls, Pythonic support
Read the full release blog: www.clarifai.com/blog/benchmarking-gpt-oss-across-h100s-a… -
Clarifai
@clarifai
Building intelligent, scalable AI agents for the enterprise is a challenge, but it's one you don't have to solve alone!
Clarifai has partnered with @RelevanceAI_ to bring you an exclusive webinar that provides a blueprint for success.
Join Clarifai’s Minh Tran and Relevance AI’s Ahmed Alassafi on August 20th for a deep dive into building Enterprise-Grade Agentic AI.
Register here: www.clarifai.com/webinar-enterprise-grade-agentic-ai -
Clarifai
@clarifai
Run your own AI coding agent locally with GPT-OSS and OpenHands!
Keep your code private while getting intelligent, context-aware coding help that can work across your entire repository!
Full tutorial: www.clarifai.com/blog/run-your-own-ai-coding-agent-locall… -
Clarifai
@clarifai
Access gpt-oss-120b model via API!
Clarifai hosts the OpenAI's GPT-OSS 120B, the latest open weight reasoning model.
Try out the model here, or access it via API using the Clarifai SDK or the OpenAI-compatible API: clarifai.com/openai/chat-completion/models/gpt-oss-120b -
Clarifai
@clarifai
We are going live for Technical Deep Dive into GPT-OSS in 5 mins! 🚀
Join us here: www.youtube.com/live/ngBdQ5J0ApI
Clarifai
@clarifai
Join us today for a Live Technical Deep Dive into GPT-OSS with the Clarifai team!
We’ll explore OpenAI’s GPT-OSS-120B and GPT-OSS-20B models, diving into their architectures, performance, and accessibility. This free, open-weight milestone is set to shape the future of AI deployment.
📅 Date: Friday, 8th Aug 2025 (Today)
🕒 Time: 1 PM EDT (just a few hours away!)
Watch live here: www.youtube.com/live/ngBdQ5J0ApI -
Clarifai
@clarifai
Join us today for a Live Technical Deep Dive into GPT-OSS with the Clarifai team!
We’ll explore OpenAI’s GPT-OSS-120B and GPT-OSS-20B models, diving into their architectures, performance, and accessibility. This free, open-weight milestone is set to shape the future of AI deployment.
📅 Date: Friday, 8th Aug 2025 (Today)
🕒 Time: 1 PM EDT (just a few hours away!)
Watch live here: www.youtube.com/live/ngBdQ5J0ApI -
Clarifai
@clarifai
Day 0 support for GPT‑5 on Clarifai! 🔥
OpenAI has introduced GPT‑5, its most advanced reasoning and generative model yet. It offers performance leaps in accuracy, speed, context handling, structured thinking, and problem solving.
There are three variants: 🧵👇Post not available here, view on X
Clarifai
@clarifai
3. GPT‑5 Nano – tuned for ultra-low-latency inference, ideal for edge use cases and budget-sensitive environments
Try it out here: clarifai.com/openai/chat-completion/models/gpt-5-nano
Clarifai
@clarifai
Clarifai supports all three on the Platform. If you're on Essential, Professional, or Enterprise plans, you can:
• Try GPT‑5 models in the Playground
• Integrate them into your applications using the API -
Clarifai
@clarifai
LIVE WEBINAR. Enterprise-Grade Agentic AI: Build Faster, Scale Smarter! 🚀
Building intelligent, production-ready AI agents is hard.
From orchestrating LLMs to managing context, integrating with tools, and deploying efficiently—there’s a lot to get right.
Join Clarifai and @RelevanceAI_ to learn how to accelerate development and build agentic systems that scale.
📅 August 20 | 4 PM ET | 1 PM PT
Register here: www.clarifai.com/webinar-enterprise-grade-agentic-ai -
Clarifai
@clarifai
🚀 Clarifai now hosts GPT OSS 120B, the latest open weight reasoning model from OpenAI!
Try it out here: clarifai.com/openai/chat-completion/models/gpt-oss-120bThis media couldn't be displayed. See it on the original post. -
Clarifai
@clarifai
Run OpenAI’s latest open‑source GPT‑OSS‑20B model locally via Public API!
OpenAI has released GPT‑OSS‑20B, a powerful open‑source model for reasoning and coding tasks.
In this demo, @mattzeiler walks through how to run GPT‑OSS‑20B using Ollama and expose it through a public API using Clarifai Local Runners.This media couldn't be displayed. See it on the original post.
Clarifai
@clarifai
Check out the full documentation and examples here: docs.clarifai.com/compute/local-runners/
-
Clarifai
@clarifai
AI isn’t just about models! It’s about the infrastructure that powers them
The latest GigaOm Radar breaks down top AI platforms for real-world scale, speed & security.
Clarifai is featured among the leaders.
Read the full report: portal.gigaom.com/reprint/gigaom-radar-for-ai-infrastruct… -
Clarifai
@clarifai
🚀 Qwen3‑Coder‑30B‑A3B‑Instruct is now on Clarifai!
• 30.5B parameters for efficiency + scale.
• 48 layers for deep reasoning on complex code.
• 256K context window for full‑repo understanding.
• Top performer for agentic coding.
Try it now in the Playground or access it via API: clarifai.com/qwen/qwenCoder/models/Qwen3-Coder-30B-A3B-In… -
Clarifai
@clarifai
As enterprises move beyond basic generative AI into the era of agentic AI, the challenges are evolving!
It is no longer just about building models. It is about optimizing AI workloads, reducing hidden costs, and accelerating deployment at scale.
Our new Executive Guide shares strategies and insights to help you tackle these challenges effectively.
Read here: www.clarifai.com/ebook-clarifai-executive-guide-2025 -
Clarifai
@clarifai
Clarifai has been recognized as a "Challenger" and "Fast Mover" in the GigaOm Radar for AI Infrastructure v1.
AI agents need more than good models — they need scalable, secure, cost-optimized infrastructure. That’s what Clarifai delivers.
👉 Report here: www.clarifai.com/gigaom-ai-infrastructure-report-optimize… -
Clarifai
@clarifai
A100 vs H100
We compared the NVIDIA A100 (Ampere) and H100 (Hopper) across:
• Key architectural upgrades like FP8, Transformer Engine, HBM3, PCIe Gen5
• Training, inference, and HPC benchmarks
• Cost vs performance trade‑offs
Read the blog here:
www.clarifai.com/blog/nvidia-a100-vs.-h100-choosing-the-r… -
Clarifai
@clarifai
We built an AI Agent from scratch!
Here’s what you will learn:
• What makes an AI Agent
• Agentic design patterns (ReAct, Reflection, Multi-Agent)
• How to build one using CrewAI and Clarifai models
Detailed guide: www.clarifai.com/blog/build-an-ai-agent-from-scratch-with… -
Clarifai
@clarifai
Missed our recent research seminar? Watch the full session now.
In this talk, Rosie Zhao presents findings from a recent preprint exploring the underlying mechanisms of reinforcement learning (RL)-based finetuning in verifiable domains like mathematical reasoning and coding—a key step in post-training large language models.
Key takeaways:
• How RL finetuning amplifies patterns from pretraining data
• Scale-dependent biases in model behavior
• Transfer learning effects across datasets
Watch the seminar here:
youtu.be/SD5raqvYG-0 -
Clarifai
@clarifai
Run Ollama models locally and make them accessible via a public API using Local Runners! 🚀
Keep your models private, cut costs, and integrate them into any project — all while running on your own machine.
Detailed guide here: www.clarifai.com/blog/run-ollama-models-locally-and-make-… -
Clarifai
@clarifai
Sovereign AI isn’t optional — it’s essential.
At #raise2025, Clarifai Founder & CEO @mattzeiler joined leaders from Oracle, Tensorwave, Vaire computing, and Credo AI to discuss how data control, compliance & national strategies are reshaping AI.
🎥 Watch the panel: youtu.be/6IIfcivyy4sThis media couldn't be displayed. See it on the original post. -
Clarifai reposted
Activepieces
@activepieces
🚀 Build smarter AI workflows with @clarifai 🤝 @activepieces
We’re excited to welcome Clarifai into the Activepieces ecosystem.
Clarifai is a full-stack AI platform that helps developers and teams deploy, scale, and optimize AI models across any environment.
In a fun and insightful interview with Minh Tran, Technical Product Marketer at Clarifai, we learn more about Clarifai and the future of AI deployment
Watch the video on YouTube to learn more: www.youtube.com/watch?v=lr2v4TfA_gY
If you want a tutorial on how to use Clarifai and Activepieces to build powerful AI workflows, check out this tutorial:
www.activepieces.com/blog/building-powerful-ai-workflows-… -
Clarifai
@clarifai
Clarifai is proud to be named a “Challenger” and “Fast Mover” in the GigaOm Radar for AI Infrastructure v1 Report!
Clarifai was recognized for:
• Superior scalability, flexibility, and pricing transparency
• Rich libraries of pre‑built models and workflows
• Advanced multi‑modal capabilities for text, image, and audio
Read the press release here: www.clarifai.com/press-release-clarifai-recognized-as-a-c… -
Clarifai
@clarifai
Clarifai Research Seminar: Reinforcement Learning for Reasoning in LLMs!
Join Rosie Zhao, a third-year PhD student at Harvard, to explore how reinforcement learning improves reasoning in large language models.
Rosie will cover fine-tuning on math and coding tasks, scaling effects, generalization to new problems, and whether RL can drive models beyond what pretraining achieves.
Register here: www.clarifai.com/research-seminar-rl-based-fine-tuning -
Clarifai
@clarifai
Run any Ollama model locally and expose it with a public endpoint in just 2 commands using Clarifai’s local-runner.
Here’s how to get started 🧵Post not available here, view on X
Clarifai
@clarifai
Step 3: Run the public endpoint
After init, run this command to serve your model through Clarifai
It wraps your local model in an OpenAI-compatible Clarifai API. Requests to the endpoint are forwarded to your Ollama model running locally on your machine.
Clarifai
@clarifai
Find the complete guide here:
github.com/Clarifai/runners-examples/tree/main/local-runn…
Read more about Local Runners here: www.clarifai.com/products/local-runners -
Clarifai
@clarifai
Choosing between NVIDIA A10 and A100 for your AI workloads?
We broke down the specs, trade-offs, and how to run on either GPU without cloud lock-in.
👉 Full comparison: www.clarifai.com/blog/nvidia-a10-vs-a100-choosing-the-rig… -
Clarifai
@clarifai
Clarifai was at RAISE Summit last week. The conversations around Sovereign AI, Local Runners, and portable deployment showed just how quickly AI infrastructure is evolving.
Here’s a quick recap 👇Post not available here, view on X
Clarifai
@clarifai
At Clarifai booth, the team spoke with developers about deploying workloads on dedicated compute. They ran live demos of Local Runners and walked through real use cases.
Clarifai
@clarifai
Thanks to everyone who stopped by. Great energy, thoughtful discussions, and real momentum around building AI that works on your terms.
-
Clarifai
@clarifai
Clarifai Research Seminar – July 24 at 12 PM ET!
"Beyond Pretraining: RL-Based Fine-Tuning for Reasoning"
Harvard PhD student Rosie Zhao will share:
– RL from scratch on math and code datasets
– Scale bias and pattern amplification
– Transfer to unseen tasks
– Can RL help models learn beyond pretraining?
📍 Register here to join live: www.clarifai.com/research-seminar-rl-based-fine-tuning -
Clarifai
@clarifai
. @mattzeiler announced Clarifai’s Local Runners at RAISE2025!
Clarifai’s Local Runners offer an efficient and flexible way to run models on your own infrastructure — while still exposing a Public API to access them from anywhere.
Check out it here: clarifai.com/products/local-runners -
Clarifai
@clarifai
Clarifai now supports token-based billing for select models running on shared compute!
This gives you finer control over costs, especially for LLMs where input and output token usage can vary widely.
Models on dedicated compute will continue using time-based pricing.
Try out models in the Playground here:
clarifai.com/playground -
Clarifai
@clarifai
Build and deploy a custom MCP server from scratch with FastMCP + Clarifai!
The guide covers tool creation (search, extract, keyword research), local testing, and deployment to dedicated compute.
Works with any MCP-compatible agent.
Read the Blog here: www.clarifai.com/blog/build-and-deploy-a-custom-mcp-serve… -
Clarifai
@clarifai
Big update for AI developers! 🚀
You can now host your own tools & APIs as MCP servers on Clarifai — and call them from any AI agent.
Plus: OpenAI-compatible API now live. Use 100+ models without changing your code.
Read the Full Press Release here: www.clarifai.com/press-release-clarifai-accelerates-agent… -
Clarifai
@clarifai
AI agents are taking off — but building and deploying them responsibly is still hard!
The latest @Gartner_inc Market Guide shares practical insights on governance, guardrails, and platform capabilities.
Clarifai is listed as a Representative Vendor, offering tools to support responsible agent and multimodal app deployment.
Read the report here: www.clarifai.com/report/market-guide-for-ai-application-d… -
Clarifai
@clarifai
Build a local AI agent using Clarifai Local Runners!
• FastAPI backend (SQLite)
• MCP server with tools via FastMCP
• Expose the MCP server via Public Endpoint using Local Runners
• Connect it to a CrewAI agent
• Interact via Streamlit UI
100% Local
Check out the Video here: youtu.be/JOdtZDmCFfk -
Clarifai
@clarifai
Clarifai 11.6 is here, introducing Local Runners! 🚀
This release is a major step forward in letting you run and expose AI models directly from your own hardware as an API endpoint without uploading them to the cloud.
Key highlights of Local Runners:
• Expose your local model as an API endpoint
• Run and test models directly on your machine with full platform integration
• Ngrok for AI models with secure routing and zero infrastructure setup
Along with Local Runners, this release includes improvements across compute orchestration, token-based billing, and new models like Gemma-3n-E2B.
Read the full release blog:
👉 www.clarifai.com/blog/clarifai-11.6-introducing-local-run… -
Clarifai reposted
TechDogs Inc
@TechDogs_Inc
@clarifai unveils AI Runners, offering developers flexible options for deploying and managing AI models on local machines or private servers via a seamless API. Know more. www.techdogs.com/tech-news/pr-newswire/clarifai-unveils-a…
#technews #technology #trendingnews #explorepage #techdogs #tdnews #tddailynews #techdogsnews -
Clarifai
@clarifai
We launched Local Runners today at #raise2025 in Paris! 🎉
Local Runners are ngrok for AI models
Run models from your laptop or server, serve them via Clarifai’s cloud API
Read more about it here: www.clarifai.com/products/local-runners
Stop by Booth #8 to see it in action. -
Clarifai
@clarifai
We’re excited to launch AI Runners! 🚀
AI Runners are ngrok for AI models. Serve models from your local machine or private server through Clarifai’s public API.
Key highlights:
✅ Connect local models to Clarifai
✅ Keep data in your infra
✅ Chain with Clarifai-hosted models
✅ Build agentic AI workflows
✅ Scale to Kubernetes clusters
✅ No complex networking or vendor lock-in
🎯 Developer Plan: $1/month for the first year
Read the full release here: www.clarifai.com/press-release-clarifai-unveils-ai-runner… -
Clarifai
@clarifai
LLMs aren’t just about prompts. They’re about structure!
Roles like system, user, and assistant keep the conversation on track. Roles like tool_use and tool_result make agents reliable and capable of reasoning and action.
Full breakdown: www.clarifai.com/blog/agentic-prompt-engineering -
Clarifai reposted
Vultr
@Vultr
🧠 Scaling AI doesn’t have to mean spiraling costs.
Join @clarifai CEO Matt Zeiler at the Vultr booth at RAISE Summit:
🗓️ July 8 | 2:30 PM
🗓️ July 9 | 3:00 PM
Discover how orchestration cuts compute costs by up to 90% and helps teams deploy AI at scale, anywhere. #raisesummit -
Clarifai
@clarifai
Clarifai will be at RAISE 2025 next week! 🚀
Join us live at Booth #8 to see how we orchestrate AI workloads at any scale, on any infrastructure
Our CEO, @mattzeiler, will speak on Sovereign AI
🗓️ July 8 | ⏰ 11:00–11:40 AM CET
www.clarifai.com/events-clarifai-at-raise-event-2025 -
Clarifai
@clarifai
CEOs are bullish on AI (74% believe it'll impact their industry the most), but nearly half of AI initiatives fail to meet ROI targets! 📉
The @Gartner Market Guide for AI App Dev Platforms discusses how to bridge this gap. In our opinion, you can leverage vendors like Clarifai to drive real business value.
Get the report here: www.clarifai.com/report/market-guide-for-ai-application-d… -
Clarifai
@clarifai
🚀 Heading to #raise2025 in Paris?
Clarifai will be at Booth 8, showcasing how our Compute Orchestration lets you run AI anywhere — cloud, on-prem, or edge.
📍 July 8–9 | Le Carrousel du Louvre
🎟️ We're giving away 3 free tickets — apply here:
share.hsforms.com/15mySGIKMRP6Eo6h_fIL3cA3n10g?partner_so… -
Clarifai
@clarifai
Gemma 3n (text-only) is now available on Clarifai! 🎉
Available in two variants:
⚡ E2B for mobile-friendly tasks
⚡ E4B for full performance on larger devices
Uses PLE caching, MatFormer, and conditional param loading.
👉 clarifai.com/gcp/generate/models/gemma-3n-E2B-it-GGUF-4bi…
👉 clarifai.com/gcp/generate/models/gemma-3n-E4B-it-GGUF-4bi… -
Clarifai
@clarifai
You can now use image generation models from the Clarifai Community through the OpenAI-compatible API endpoint!
Works out of the box with models like grok-2-image using the OpenAI SDK or any client that supports the OpenAI format.
Docs here:
www.clarifai.com/blog/openai-compatible-api-image-generation -
Clarifai
@clarifai
Benchmarking the best open-source Vision-Language Models! 🚀
Gemma 3 vs. MiniCPM 2.6 vs. Qwen 2.5 VL.
We tested latency, throughput, and concurrency on L40S GPUs using Clarifai’s Compute Orchestration.
Read the full blog:👇
www.clarifai.com/blog/benchmarking-best-open-source-visio… -
Clarifai
@clarifai
Clarifai is all set for RAISE 2025! 🚀
Join us as we showcase how leading enterprises are accelerating AI adoption — with real use cases, powerful demos, and conversations that matter.
Come meet the Clarifai team and see what’s next in AI.
Book a slot with our team here: www.clarifai.com/events-clarifai-at-raise-event-2025
#raise2025 -
Clarifai
@clarifai
AI development is tough!
Over 75% of engineering leaders struggle to integrate AI capabilities effectively.
The @Gartner Market Guide for AI App Dev Platforms offers insights to conquer these pain points & unlock true business impact.
Get the report: www.clarifai.com/report/market-guide-for-ai-application-d… -
Clarifai
@clarifai
Last day at #hpediscover 2025! ⚡
If you're around and want to meet the team, come visit us at Booth #1028.
We’re showcasing how Clarifai is rethinking AI infrastructure from the ground up — with seamless orchestration across cloud, on-prem, and air-gapped environments.
📍 Booth #1028
📅 June 26 -
Clarifai
@clarifai
Run inference on Clarifai models using the Vercel AI SDK!
The AI SDK supports the OpenAI-compatible API.
Below is a simple example of how to access the model from the Community via Clarifai’s OpenAI-compatible endpoint.
Find the example here: docs.clarifai.com/compute/providers/vercel -
Clarifai
@clarifai
Day 2 at #hpediscover 2025! ⚡
We’re live in Vegas at Booth 1028.
Stop by to see how Clarifai is rethinking AI infra across cloud, on-prem, and air-gapped setups — with smarter orchestration and a unified control plane.
Book some time with the team: www.clarifai.com/events-clarifai-at-hpe-discover-event-2025 -
Clarifai
@clarifai
Streaming Clarifai models using the OpenAI-compatible API with @LiteLLM ⚡
Clarifai now supports real-time token streaming for any hosted model. Just set 𝐬𝐭𝐫𝐞𝐚𝐦=𝐓𝐫𝐮𝐞 and you're good to go.
Below is an example snippet that shows how simple it is with litellm:
Clarifai
@clarifai
Find the docs here:
docs.clarifai.com/compute/providers/litellm -
Clarifai
@clarifai
Inference with the LiteLLM Python SDK!
You can now run models from Clarifai Platform with minimal setup using LiteLLM.
Check out the code here:
docs.clarifai.com/compute/providers/litellm -
Clarifai
@clarifai
Agentic LLM for software engineering tasks! 🔥
Introducing Devstral-Small-2505
The 4-bit GGUF version of Devstral-Small-2505 is now live in the community.
Try it out in the Playground here:
clarifai.com/playground?model=Devstral-Small-2505_gguf-4b… -
Clarifai
@clarifai
Tool Calling now works with Clarifai’s OpenAI-compatible API.
Let your LLMs decide when to call APIs or custom tools based on user input. Here’s a quick weather API example to get started.
Read more here:
docs.clarifai.com/compute/providers/open-ai -
Clarifai
@clarifai
How does gradient descent really work in deep learning?
Join Jeremy Cohen (Flatiron Institute) for a live seminar on why classic optimization theory falls short—and what third-order analysis reveals instead.
🗓️ June 20 | 12 PM ET
🔗 www.clarifai.com/research-seminar-how-does-gradient-desce… -
Clarifai
@clarifai
Clarifai at @HPE Discover 2025!
June 23 to 26 in Las Vegas
Visit our booth to see how we're scaling AI infrastructure across cloud, on-prem, and air-gapped environments.
Book a meeting with the team here:
www.clarifai.com/events-clarifai-at-hpe-discover-event-2025
#hpediscover -
Clarifai
@clarifai
Build MCP-Powered AI Agents using Clarifai!
In this hands-on session, we have walked through:
🛠️ Creating a custom MCP server with tools using Fast MCP
🚀 Deploying it to dedicated compute on the Clarifai Platform
🤖 Building an AI Blog Writing Agent using the deployed MCP server with Clarifai and Crew AI
If you missed it during Major League Hacking(MLH) GenAI Hack Week, you can now watch the full recording here: 👇
www.youtube.com/watch?v=7L_BxQLQb3w&t=1905s -
Clarifai
@clarifai
Clarifai is now OpenAI-Compatible!
Use Clarifai-hosted models with the openai Python client. No rewrites needed.
Just update your base URL to Clarifai’s OpenAI-compatible endpoint.
Run chat completions, stream responses, and integrate tools using your existing OpenAI code.
Check out the examples here: 👇docs.clarifai.com/compute/models/inference/open-ai -
Clarifai
@clarifai
Clarifai 11.5 is live — now with support for AI Agents and Model Context Protocol (MCP)!
Build autonomous Agents, deploy custom MCP servers, and use the OpenAI-compatible API.
New models + SDK upgrades too.
Read more: clarifai.com/blog/clarifai-11.5-introducing-support-for-a… -
Clarifai
@clarifai
💡 Dev Tip — Clean up deployments fast with the Nodepool API!
Use delete_deployments to remove multiple deployments at once. Just fetch the IDs and pass them in a single call.
Quick, simple, efficient.
Check out the code here: docs.clarifai.com/compute/deployments/manage-compute/ -
Clarifai
@clarifai
Clarifai is heading to @AMD Advancing AI 2025! 🚀
📍 Santa Clara, CA | 📅 June 12
We’ll be showcasing how Clarifai simplifies AI infrastructure by unifying compute across cloud, on-prem, and edge.
Come meet the team, Book a meeting here: 👇
www.clarifai.com/events-clarifai-at-amd-advancing-ai-2025 -
Clarifai
@clarifai
Clarifai is at AWS Summit Washington, DC!
Join our team members Jonathan Padgett and Dan Stroman at the Walter E. Washington Convention Center
If you’re attending, stop by and say hi — we’d love to connect.
#awssummit #publicsectorai #clarifai -
Clarifai
@clarifai
We're teaming up with @MLHacks for a hackathon: Build AI Agents & custom MCPs! 🚀
This week-long hackathon is your chance to learn and build AI agents and custom Model Context Protocol Servers (MCP) using Clarifai Platform
🗓️ June 13-19
Register here: ghw.mlh.io/events/generative-ai -
Clarifai
@clarifai
Clarifai is heading to HPE Discover 2025! 🚀
Meet Jonathan Padgett (VP of Sales) and Dan Stroman (Global Channel Partner) at our booth to explore the future of AI infrastructure at scale.
📍 June 23–26 | Las Vegas
Book a meeting with the team here: 👇
www.clarifai.com/events-clarifai-at-hpe-discover-event-2025 -
Clarifai
@clarifai
Access DeepSeek-R1-0528-Qwen3-8B with an API!
This new model builds on DeepSeek-R1 v2 with major improvements in reasoning and logic, powered by more efficient computation and optimization.
Access the model via the Clarifai's OpenAI-compatible API endpoint!
Clarifai
@clarifai
Checkout the model here:
clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-R1… -
Clarifai
@clarifai
Marketing Agent Crew 🎯
We have built an AI marketing team Agent to generate and analyze marketing strategies
Built with:
• CrewAI for multi-agent orchestration
• gpt-4o model from the Clarifai community
Check out the example below: 👇
github.com/Clarifai/examples/tree/main/agents/CrewAI/mark… -
Clarifai
@clarifai
Clarifai at AMD Advancing AI 2025 — June 12 in Santa Clara! 🚀
Meet CEO @mattzeiler and Head of GTM Sajai Krishnan as we showcase our unified AI compute platform.
Book a meeting: www.clarifai.com/events-clarifai-at-amd-advancing-ai-2025 -
Clarifai
@clarifai
AI Blog Writing Agent using Clarifai and CrewAI! 🚀
In this video, we build an AI-powered blog writing agent that generates complete blog posts from scratch.
We use:
• CrewAI to manage agent orchestration
• Gemini 2.5 Pro model powered by Clarifai
• Streamlit to create a simple and interactive UI
👉 Checkout the Video: youtu.be/1XZT2wD4UzQ -
Clarifai
@clarifai
We’ve improved the Clarifai Playground! 🚀
Now, the selected deployment ID is automatically added to the inference code. This makes testing models faster and easier.
Playground is your quick way to try out powerful AI models.
Try it out → clarifai.com/playground -
Clarifai
@clarifai
We’re hosting a San Francisco hackathon focused on building AI agents with Clarifai + Model Context Protocol (MCP) 🚀
🎯 Theme: Build agents w/ reasoning, memory & tools using Clarifai
📍 May 30 · SF
Build agentic apps with Clarifai → lu.ma/mcphack -
Clarifai
@clarifai
Clarifai just won the 2025 AI TechAward for Best AI Development Framework. The award recognizes our Compute Orchestration platform! 🔥
Run any model on any compute with full control and flexibility.
Read the press release: www.clarifai.com/press-release-clarifai-wins-2025-ai-tech… -
Clarifai
@clarifai
Clarifai is heading to Santa Clara for the AI & Big Data Expo / TechEx North America!
We're excited to be at @Vultr's booth showing how you can orchestrate AI workloads at any scale.
Don’t miss Sajai Krishnan, our Head of GTM, speaking on June 5th at 12:30PM PT!
techexevent.com/northamerica/ -
Clarifai
@clarifai
Missed the Research Seminar?
Stella Yu shared insights on Unscripted Grounded Visual Learning, showing how vision can emerge from raw images and videos without labels or semantics.
From CLIP to how infants perceive the world.
👉 Watch it here youtu.be/4lpZRBSN_cI -
Clarifai
@clarifai
Join us in San Francisco next Friday, May 30 for the MCP & Agents Hackathon
💡 Build your own agent applications powered by Clarifai - participants get free credits and hands-on support throughout the event
🛠️ Create custom MCPs and deploy them to the Clarifai platform to host and serve directly from the cloud
🚀 Tackle real-world challenges using multimodal agents with advanced reasoning and tool use
RSVP here: lu.ma/mcphack -
Clarifai
@clarifai
Missed our live webinar with @HPE?
We explored how AI is driving operational efficiency across industries, from real-time pattern detection to intelligent automation.
Watch it on demand: 👇
www.clarifai.com/webinar-boosting-boosting-business-opera… -
Clarifai
@clarifai
Research Seminar: Unscripted Grounded Visual Learning! 🚀
We’re hosting Prof. Stella X. Yu (UMich) to explore how visual learning can happen without labels or language.
🗓️ 22nd May, 12pm ET
RSVP here: www.clarifai.com/unscripted-grounded-visual-learning -
Clarifai
@clarifai
🚀 We are excited to share that Clarifai joins @Vultr Cloud Alliance to deliver scalable, cost-optimized full-stack AI.
Run any model on any GPU across cloud, edge, or hybrid. Save at least 70% on A100 GPUs compared to hyperscalers.
Read the press release: www.clarifai.com/press-release/clarifai-vultr-cloud-alliance
#clarifai #vultr -
Clarifai
@clarifai
🚀 Set up your own dedicated compute!
Learn how to set up dedicated compute on Clarifai and deploy models using your preferred cloud provider.
The video walks through creating a cluster, configuring a node pool, choosing instance types, and deploying a community model step by step, with scaling options and cost optimizations included.
Watch it here: youtu.be/WZkcEQ1D7x0 -
Clarifai
@clarifai
Clarifai 11.4 is live 🎉
Python-first model upload & inference
• New community models: Llama-4-Scout, Qwen3, o3, o4-mini
• Upgrades to Playground, Control Center & SDKs
• Build faster with less code.
Read the Full Release Blog 👇
www.clarifai.com/blog/clarifai-11.4-faster-model-deployme… -
Clarifai
@clarifai
Access Qwen3 with just a few lines of Python code! 🔥
Qwen3-30B-A3B is a powerful Mixture-of-Experts model designed for advanced reasoning, coding, instruction following, and multilingual tasks.
Now available on the Clarifai Platform with full API access
Here’s how you can access it using the Clarifai Python SDK.
Try it out here: www.clarifai.com/explore/Qwen3-30B-A3B-GGUF -
Clarifai
@clarifai
What exactly is a GPU cluster and why does it matter for GenAI?
We just published a deep dive on how GPU clusters power scalable, high-performance AI.
Covers:
• What makes a GPU cluster
• How GPU fractioning reduces cost
• Deploying with Clarifai Compute
• Picking the right GPU (T4, L40S, A100)
Read: www.clarifai.com/blog/what-are-gpu-clusters-and-how-they-… -
Clarifai
@clarifai
Starting in 1 hour! 🚀
Clarifai and @HPE are going live to share how AI is transforming manufacturing and industrial operations, from defect detection with vision models to decision acceleration with LLMs and automation using agents.
👉 Join us here: www.clarifai.com/webinar-boosting-boosting-business-opera… -
Clarifai
@clarifai
🚀 Moving from AI experiments to real-world operations!
Clarifai and @HPE are hosting a session tomorrow on how AI is solving practical problems in manufacturing and related industries.
We’ll share examples like spotting defects with vision models, using LLMs to speed up decisions, and automating repetitive tasks with agent-based systems.
🗓️ May 14 | 12 PM ET | 9 AM PT
👉 Register here: www.clarifai.com/webinar-boosting-boosting-business-opera… -
Clarifai reposted
DataSummit
@DBTADataSummit
Alfredo Ramos, @clarifai, explores how compute orchestration is optimizing AI workloads at #datasummit. Register with code DS25 today! secure.infotoday.com/RegForms/DataSummit/
-
Clarifai
@clarifai
Using OpenAI’s o3 or o4-mini in Clarifai Playground?
Adjust the Reasoning Effort setting:
• Low = fast, minimal thinking
• Medium = balanced
• High = deeper reasoning, slower
Tweak it based on whether you want speed or depth.
Try it out: clarifai.com/playground?model=o4-mini__d378fd80c0d84588a4… -
Clarifai
@clarifai
🚀 Clarifai is at #geoint2025!
Visit us at Booth #1814 in the @Carahsoft Partner Pavilion to see how we’re driving geospatial intelligence with AI.
Meet the Clarifai team to learn how we help defense & intel agencies turn unstructured data into actionable insights.
📅 Book a meeting: www.clarifai.com/events-clarifai-at-geoint-2025 -
Clarifai
@clarifai
🚀 Clarifai and @Carahsoft are hosting a webinar on AI-powered labeling & auto-annotation for the Intelligence Community.
🗓 May 21 | 2PM ET
Learn how to automate data workflows and speed up production AI.
🔗 Register: carahevents.carahsoft.com/Event/Register/610334-clarifai -
Clarifai
@clarifai
Explore some of the latest open-source and proprietary AI models now available on Clarifai — including DeepSeek, LLaMA 4, o4-mini, and more!
This quick walkthrough covers how to access trending models, run them directly in the Playground, and call them via API — no setup needed.
👉 Watch the demo here:
youtu.be/8oWjmB3Bmqk -
Clarifai
@clarifai
Clarifai is teaming up with @HPE for a live webinar on how AI is transforming ops across industries like manufacturing.
Real-time anomaly detection, automation, faster decisions — all powered by AI.
📅 May 14 | 12 PM ET
Register here: www.clarifai.com/webinar-boosting-boosting-business-opera… -
Clarifai
@clarifai
MCP vs A2A Clearly Explained!
- MCP connects LLMs to tools, data & environments.
- A2A connects AI agents for discovery & collaboration.
We wrote a detailed blog comparing MCP and A2A.
Read it here: clarifai.com/blog/mcp-vs-a2a-clearly-explained -
Clarifai
@clarifai
Qwen3-30B-A3B (MoE) is now live on Clarifai!🔥
Smaller than QwQ-32B but outperforms it in coding and reasoning tasks.
Open-weight, fast, and efficient — try it now:
clarifai.com/qwen/qwenLM/models/Qwen3-30B-A3B-GGUF
#qwen3 -
Clarifai
@clarifai
🚀 Clarifai is heading to #geoint2025!
Visit Booth #1814 in @Carahsoft Partner Pavilion to see how we turn unstructured data into insights for defense and intel missions.
Meet the team and talk with us about how to scale your AI workloads efficiently.
www.clarifai.com/events-clarifai-at-geoint-2025 -
Clarifai
@clarifai
Clarifai will be at SOF Tampa 2025 next week! 🚀
Explore how our unified AI compute platform supports scalable, mission-ready AI across defense environments.
📅 Book a meeting with the team here: www.clarifai.com/events-clarifai-at-sof-tampa-event-2025 -
Clarifai
@clarifai
🚀 Join @clarifai & @HPE for a webinar on transforming business operations with AI!
Explore how AI-powered computer vision, LLMs, and RAG can optimize processes, reduce costs, and improve decision-making, unlocking true efficiency.
📅 May 14 | 12 PM ET / 9 AM PT
🔗 Register: www.clarifai.com/webinar-boosting-boosting-business-opera… -
Clarifai
@clarifai
Live Webinar: Boost manufacturing quality control with AI-powered computer vision 🚀
We’ve partnered with @HPE to show how AI can improve safety, cut costs, and boost efficiency in production lines.
📅 May 14 | 12 PM ET / 9 AM PT
🔗 Register: www.clarifai.com/webinar-boosting-manufacturing-quality-c… -
Clarifai
@clarifai
Let's Deploy Llama 4 Scout to your own dedicated compute using Compute Orchestration:
Post not available here, view on X
Clarifai
@clarifai
Step 3: Deploy Llama 4 Scout
Llama 4 Scout is already available in the Clarifai Community.
You can deploy it directly to your dedicated compute — meaning the cluster and node pool you have just configured.
Clarifai
@clarifai
Check out the docs here to get started:
docs.clarifai.com/compute/deployments/
Try out Llama 4 Scout model here:
clarifai.com/meta/Llama-4/models/Llama-4-Scout-17B-16E-In… -
Clarifai
@clarifai
Explore OpenAI’s latest reasoning models, o3 and o4-mini, in the Clarifai Playground! 🔥
Battle test them alongside other open-source models like DeepSeek-R1 or Gemma-3B to assess performance and capabilities.
Try it out here: clarifai.com/playground?model=o4-mini__d378fd80c0d84588a4… -
Clarifai
@clarifai
Excited to share that Clarifai has been awarded a contract by the U.S. Air Force Research Lab to advance AI-powered full motion video analysis! 🎯
Partnering with the Texas Defense Innovation Forum to bring real-time AI to the edge.
Press Release: www.clarifai.com/press-release/clarifai-awarded-air-force… -
Clarifai
@clarifai
🚀 Clarifai is teaming up with @Carahsoft to host a webinar on Labeling & Auto-Annotation for Production AI in the IC.
🗓 Date: Wednesday, May 21, 2025
⏰ Time: 2:00pm ET | 11:00am PT
In this session, we'll dive into how AI can help automate data labeling and annotation to power your production AI workflows. Learn from experts how AI can accelerate your projects while saving time and resources.
Register below: carahevents.carahsoft.com/Event/Register/610334-clarifai -
Clarifai
@clarifai
Proud to see Clarifai win a 2025 AI TechAward for Compute Orchestration! 🏆
Our infra- and vendor-agnostic stack gives developers full control over how & where they run their models.
Thanks to @DevNetwork_ for the recognition!
Onward 🚀 -
Clarifai
@clarifai
Access Llama 4 Scout with an API!
Access the instruction-tuned variant of Llama-4 Scout via the API and integrate it into your application with just a few lines of code.
Use the Clarifai Python SDK to get started quickly.
Try it out the model here: clarifai.com/meta/Llama-4/models/Llama-4-Scout-17B-16E-In… -
Clarifai
@clarifai
🚀 Clarifai at Menlo CEO Summit – Starts Today!
April 23-24 | Healdsburg, CA
@mattzeiler, CEO of Clarifai, is live at the Menlo CO Summit! Ready to connect and showcase how we're transforming AI infrastructure.
You can schedule some time with Matt here: www.clarifai.com/events-clarifai-at-menlo-ceo-summit-2025 -
Clarifai
@clarifai
🔐 Want full visibility into your AI workflows?
Clarifai makes it easy with Audit Logging. You can track model usage, user activity, and system changes to strengthen security and support compliance.
Read the detailed blog below:
www.clarifai.com/blog/complete-guide-to-audit-logging-wit… -
Clarifai
@clarifai
OpenAI’s latest models: o3 and o4-mini are now live on Clarifai! 🚀
🧠 o3 excels at deep, multi-step reasoning (math, code, visual tasks)
⚡ o4-mini is fast, cost-effective, and highly accurate
Use in the Playground or access via API
Try them out 👉 clarifai.com/playground -
Clarifai
@clarifai
🚀 Missed our webinar with @Carahsoft on supercharging AI workloads? Catch the replay to learn how to:
✅ Optimize AI infrastructure
✅ Cut costs with orchestration
✅ Boost efficiency with the right tools
Access the recording here: 👇
www.carahsoft.com/learn/event/66319 -
Clarifai
@clarifai
Clarifai hosted Stanford PhD Kanishk Gandhi for a live talk on why some LLMs get better with reinforcement learning while others plateau.
He shared 4 key cognitive behaviors that drive self-improvement in reasoning systems.
Watch the full talk 👇
youtu.be/p5FcFApZwrY -
Clarifai
@clarifai
🚀 Clarifai at Menlo CEO Summit 2025 (April 23-24, Healdsburg, CA)
Want to optimize your AI workflows?
Book a meeting with @mattzeiler to see how compute orchestration can reduce costs, deploy models at scale, and leverage GPUs/CPUs.
www.clarifai.com/events-clarifai-at-menlo-ceo-summit-2025 -
Clarifai
@clarifai
Vendor lock-in makes it harder to build and manage real AI workflows.
It limits control, flexibility, and slows down iteration!
With Clarifai’s Compute Orchestration, you get the freedom to deploy your own dedicated models on any cloudincluding AWS, GCP, Oracle, Vultr, and more coming soon.
Run models on your own terms, scale as you need, and keep full control over your infrastructure.
Want to deploy on your own infra?
Check out Compute Orchestration: www.clarifai.com/products/compute-orchestration -
Clarifai reposted
Clarifai
@clarifai
Run Multiple AI Workloads on a Single GPU: GPU Fractioning Clearly Explained 🧵
-
Clarifai reposted
Artjom Kurapov
@tot_ra
Wrote MCP server for @clarifai
Let me know if you find it useful and want more features like search, annotations, manipulating models just by talking to your AI agent 🚀
github.com/tot-ra/clarifai-mcp-server-local
www.youtube.com/watch?v=aSuJxq1txm0 -
Clarifai
@clarifai
Run Multiple AI Workloads on a Single GPU: GPU Fractioning Clearly Explained 🧵
Post not available here, view on X
Clarifai
@clarifai
A major challenge in GPU fractioning is managing clusters, GPU slices, and scaling workloads.
Clarifai’s Compute Orchestration handles it all in the background—no manual tuning or infrastructure management.
Just build and ship your models.
Clarifai
@clarifai
Read more about GPU Fractioning here:
www.clarifai.com/blog/gpu-fractioning-explained-how-to-ru… -
Clarifai
@clarifai
🚀 Big news! Llama 4 Scout is now available on Clarifai.
🧠 Llama 4 Scout supports a native 10M-token context window (with 8k tokens currently supported on Clarifai), making it ideal for:
• Multi-document analysis
• Complex codebase understanding
• Personalized, intelligent workflows
It’s state-of-the-art in its size class—efficient, capable, and ready to build with.
👉 Try out Llama 4 Scout on Clarifai:
clarifai.com/meta/Llama-4/models/Llama-4-Scout-17B-16E-In… -
Clarifai
@clarifai
Clarifai is featured by AIM Research as a top AI startup transforming U.S. defense! 🚀
From cybersecurity to logistics, our on-premise AI platform powers faster decisions with custom CV and genAI.
Thanks to AIM Research for the recognition!
aimresearch.co/ai-startups/here-are-the-top-ai-startups-s… -
Clarifai
@clarifai
Clarifai 11.3 is live 🚀
Test top AI models in the new AI Playground and deploy them to your dedicated compute in just a few clicks. No setup, no hassle.
Plus updates to Control Center, platform, and SDKs.
Read the full release blog. 👇
www.clarifai.com/blog/clarifai-11.3-ai-playground-your-ll… -
Clarifai
@clarifai
Clarifai is featured in FeedSpot’s Top 100 AI Blogs!
We're proud to be recognized as one of the leading voices in the AI space. Big thanks to our readers and community for the continued support!
👉 Check out the full list here: bloggers.feedspot.com/ai_blogs/ -
Clarifai
@clarifai
We’re live at #googlecloudnext 2025! 🎉
Meet Clarifai in Las Vegas this week and discover how we’re helping teams:
✅ Optimize compute + cut infrastructure costs
✅ Deploy any AI model on any cloud
✅ Manage all AI workloads from one place
👋 Book a time with Alfredo Ramos (SVP of Platform) and the team:
clarifai.com/events-clarifai-at-google-next-event-2025 -
Clarifai
@clarifai
What makes an LLM a better reasoner? 🤔
Join us this Thursday for the Clarifai Research Seminar with Kanishk Veerdhaval Gandhi (Stanford CS PhD) as he breaks down four key behaviors that help models get smarter — even when they're wrong.
🚀 Register here: -
Clarifai
@clarifai
Upload Custom Models and deploy them to your dedicated compute using the Python SDK! 🔥
The Clarifai Platform makes it easier than ever to integrate your own AI models into production.
Whether you’re using a pre-trained model from Hugging Face or one you’ve built from scratch — the Python SDK lets you upload and deploy models with just a few lines of code.
Want to see how it works? Check out the step-by-step tutorial:This media couldn't be displayed. See it on the original post. -
Clarifai
@clarifai
Clarifai Research Seminar — April 9 🚀
Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs
Featuring Stanford PhD student Kanishk Gandhi
Learn how reasoning patterns like verification, backtracking & subgoals shape RL dynamics in LLMs — and why some models improve while others plateau.
RSVP here: www.clarifai.com/cognitive-behaviors-that-enable-self-imp… -
Clarifai
@clarifai
Clarifai at Aberdeen AI/ML Technology Exchange & Expo
📅 April 9 | Belcamp, MD
Meet our team to explore how to automate and optimize AI workflows across cloud, edge & classified environments — all from one platform.
📩 Book a meeting: www.clarifai.com/events-clarifai-at-aberdeen-ai-ml-techno… -
Clarifai
@clarifai
Clarifai at Google Cloud Next 2025!
Learn how you can:
✅ Optimize compute resources and reduce costs
✅ Deploy any AI model on any compute, at any scale
✅ Manage AI workloads with a single control plane
📅 April 9-11 | Mandalay Bay, Las Vegas
👉 Book a Meeting with the team here: www.clarifai.com/events-clarifai-at-google-next-event-2025 -
Clarifai
@clarifai
🚀 Introducing AI Playground!
Test the latest AI models like DeepSeek-R1, Gemma-3, and more—all in one place, no setup needed.
• Run inference without authentication
• Compare models & results
• Get ready-to-use API snippets
Try it now 👉 clarifai.com/playground -
Clarifai
@clarifai
Any model. Any GPU. Any infrastructure!
Join us tomorrow for an exclusive webinar with Carahsoft, where we’ll dive into:
✅ Enhancing AI infrastructure for top performance
✅ Cutting costs with compute orchestration
✅ Improving efficiency with a complete AI toolkit
📅 Date: Wednesday, April 2, 2025
⏰ Time: 2:00 PM ET | 11:00 AM PT
Register here: carahevents.carahsoft.com/Event/Register/583976-cs1 -
Clarifai
@clarifai
Join us next week for an exclusive webinar with @Carahsoft, where we’ll explore how to:
✅ Optimize AI infrastructure for peak performance
✅ Reduce costs with compute orchestration
✅ Boost efficiency with a complete AI toolbox
📅 Date: Wednesday, April 2, 2025
⏰ Time: 2:00 PM ET | 11:00 AM PT
Register here: carahevents.carahsoft.com/Event/Register/583976-cs1?auth=… -
Clarifai
@clarifai
🚀 Clarifai at Aberdeen AI/ML Technology Exchange & Expo: Unifying AI Compute: One Platform, Endless Possibilities
Join Joshua Wolfe, Public Sector Account Executive, as he discusses how Clarifai's Compute Orchestration enables seamless automation and optimization of AI model training, deployment, and inferencing across distributed environments.
If you're attending and your team is looking for ways to efficiently optimize AI workflows, let’s connect.
📅 April 9 | BELCAMP, MARYLAND
Find time here: www.clarifai.com/events-clarifai-at-aberdeen-ai-ml-techno… -
Clarifai
@clarifai
Deploy Models on Dedicated Compute, Right from the New UI! 🚀
The Clarifai Platform just got a major upgrade! Our revamped UI makes deploying models to dedicated compute faster and easier than ever.
Watch the full walkthrough video below! 👇This media couldn't be displayed. See it on the original post. -
Clarifai
@clarifai
Clarifai at Google Cloud Next 2025!
Meet Alfredo Ramos (SVP of Platform) and Rachael Kloek (Account Executive) to see how we’re unifying AI compute in one powerful platform.
Learn how you can:
✅ Optimize compute resources and reduce costs
✅ Deploy any AI model on any compute, at any scale
✅ Manage AI workloads with a single control plane
✅ Leverage GPUs, CPUs, TPUs, and more
Power your AI workflows with a production-ready deep learning platform built for developers, data scientists, and ML/MLOps teams.
📅 April 9-11 | Mandalay Bay, Las Vegas
👉 Book a Meeting with the team here: www.clarifai.com/events-clarifai-at-google-next-event-2025 -
Clarifai
@clarifai
Clarifai Research Seminar: Symmetry in Machine Learning and Why It Matters! 🚀
Symmetry plays a fundamental role in science, helping simplify complex systems. In deep learning, symmetry appears in data, model architectures, and network parameters, often with practical implications.
Henry Kvinge from Pacific Northwest National Laboratory and the University of Washington shared how symmetry shapes machine learning models.
If you missed the session, watch it here: youtu.be/ZMgM0klASL8 -
Clarifai
@clarifai
Choosing between NVIDIA A10 and L40S for AI workloads?
Here's a breakdown of specs, benchmarks, and cost comparisons focused on LLM training, fine-tuning, and generative tasks.
We also cover pricing details and how Clarifai Compute Orchestration helps you run on either GPU across clouds with no vendor lock-in.
Full comparison:👇
www.clarifai.com/blog/nvidia-a10-vs-l40s-gpus-for-ai-work…
#ai #gpu -
Clarifai
@clarifai
Access Gemma-3 via API with just a few lines of code! 🔥
Gemma-3 is a lightweight, open model with vision understanding, multilingual support, and extended context windows (up to 128K tokens).
The instruction-tuned Gemma-3-4B model is now available on the Clarifai Platform.
Here’s how you can access it using the Clarifai Python SDK.
Try it out: clarifai.com/gcp/generate/models/gemma-3-4b-it -
Clarifai
@clarifai
Introducing the New Playground to explore the latest AI models! 🔥
We’ve completely revamped the Clarifai Platform experience. You can now easily select any model and start testing or deploy it directly to your dedicated compute. Just head to the Home Page: clarifai.com/home to explore.
We’ve also launched the Playground: clarifai.com/playground — a space where you can test and explore the latest LLMs and AI models all in one place. No setup needed. Switch between models and try them out instantly.
Plus, you’ll find ready-to-use API code snippets to integrate these models into your projects within minutes.
Check it out now and experience the new Playground: clarifai.com/playground -
Clarifai
@clarifai
We’ve built an OCR app powered by Google’s Gemma 3!
It extracts text from images and outputs it as structured Markdown.
The code is open-source.
Try it out here: clarifai.com/sumanth/AI-Apps/modules/gemma_ocr -
Clarifai
@clarifai
Maximize Your Compute Potential for Mission-Critical AI Projects! 🚀
Join us for an exclusive webinar with @Carahsoft, where we’ll dive into how to:
✅ Optimize your AI infrastructure
✅ Reduce costs with compute orchestration
✅ Boost efficiency with a complete AI toolbox
📅 Date: Wednesday, April 2, 2025
⏰ Time: 2:00 PM ET | 11:00 AM PT
👉 Register here: carahevents.carahsoft.com/Event/Register/583976-cs1?auth=… -
Clarifai
@clarifai
We are excited to present Research Seminars!
Join us this Thursday at 12 PM EST for the next Clarifai Research Seminar featuring Harbir Antil, Head of the George Mason University Center for Math and AI.
Harbir will dive into Optimization and High-Fidelity Digital Twins, a must-attend for anyone interested in AI, math, and digital simulations!
📆Date: Thursday
⏰Time: 12 PM EST
See you there!
Add to your calendar here: accounts.google.com/v3/signin/identifier?continue=https:/… -
Clarifai
@clarifai
Clarifai will be at the 2025 Artificial Intelligence Summit 🚀
Join Alfredo Ramos, SVP of Platform, and Joshua Wolfe, Public Sector Account Executive, at the 2025 Artificial Intelligence Summit this week on March 20 in McLean, VA at Jones Branch Drive.
If you're attending, book a meeting with the team to discover how to optimize and orchestrate your AI workloads:
👉 www.clarifai.com/events/clarifai-at-artificial-intelligen… -
Clarifai
@clarifai
Heading to #gtc25 today?
Join Matthew Zeiler, CEO & Founder of Clarifai, and the Clarifai team in San Jose, CA🚀
Book a meeting with the team to discover how to optimize and orchestrate your AI workloads:
👉 www.clarifai.com/events-clarifai-at-nvidia-ai-summit-2025
#gtc25 -
Clarifai
@clarifai
Orchestrating AI workloads at scale? Let’s make it simple.
Join @mattzeiler at #gtc25 in the @Vultr Theater Session as he shares how Clarifai’s Compute Orchestration Layer helps enterprises cut compute costs by up to 90% while seamlessly scaling AI.
Booth 1533 | March 18 | 3:00 PM & March 19 | 6:00 PM
Let’s optimize AI infrastructure!
#ai #cloudcomputing #mlops #gtc25 #clarifai #vultr -
Clarifai
@clarifai
🚀 What is Data Labeling? The Key to Building High-Quality AI Models
Data is the fuel that powers AI, but without accurate labeling, even the best models can struggle. Learn how data labeling transforms raw data into AI-ready insights and discover the best methods for ensuring accuracy and efficiency.
In this blog, we break down:
✅ What data labeling is and why it’s mission-critical
✅ Different labeling methods and how to choose the right one
✅ How Clarifai simplifies the entire process with AI-powered tools
👉Read here: www.clarifai.com/blog/what-is-data-labeling-the-key-to-bu… -
Clarifai
@clarifai
🚀 Clarifai 11.2 is here!
We’re excited to introduce the latest updates in Clarifai 11.2, making AI development even more efficient with automated data labeling at scale with Human-in-the-Loop and powerful platform enhancements.
- Gemma-3 now on Clarifai – The latest open model from Google DeepMind is available on the platform.
- Control Center updates – Improved visibility and management across your AI workflows.
- Data Utils library enhancements – Streamlined data processing for faster, more efficient AI model training.
- Python SDK improvements – Expanded functionality and optimizations for a smoother developer experience.
Read the release blog here: www.clarifai.com/blog/clarifai-11.2-automate-data-labelin… -
Clarifai
@clarifai
🚀 Clarifai is Now Available on Amazon Web Services (AWS) Marketplace! 🔥
We’re excited to announce that Clarifai’s AI Orchestration Platform is now accessible through AWS Marketplace, making it easier than ever for AWS customers to deploy, scale, and optimize AI models.
With Clarifai on AWS, you get:
✅ Simplified AI Deployment – Easily manage AI models & workflows within AWS infrastructure
✅ Optimized Compute – Reduce costs with autoscaling & flexible compute options
✅ Full-Stack AI Capabilities – From data labeling to model training & inference
✅ Enterprise-Grade Security – AI you can trust, built for compliance
Read the Press Release here: www.clarifai.com/press-release/clarifai-announces-amazon-… -
Clarifai
@clarifai
Clarifai at HPE Tech Jam 2025 🚀
Join Joshua Wolfe, Public Sector Account Executive, and Teddy Oswari, Senior Solution Engineer, on March 24–28 in Orlando, Florida.
If you're attending, book a meeting with the team to discover how to optimize and orchestrate your AI workloads:
👉 www.clarifai.com/events/clarifai-at-artificial-intelligen…
#hpetechjam2025 -
Clarifai reposted
Clarifai
@clarifai
Gemma-3 is now on Clarifai!🔥
Congrats to the Google DeepMind team on the release of Gemma-3, the best open, non-reasoning model!
Gemma-3 is a lightweight open model family (1B–27B parameters) with vision understanding, multilingual support, and extended context windows (up to 128K).
The instruction-tuned Gemma-3-4B model is now available on the Clarifai Platform. Try it out for free!
👉 Gemma-3-4B-it: clarifai.com/gcp/generate/models/gemma-3-4b-it -
Clarifai
@clarifai
Gemma-3 is now on Clarifai!🔥
Congrats to the Google DeepMind team on the release of Gemma-3, the best open, non-reasoning model!
Gemma-3 is a lightweight open model family (1B–27B parameters) with vision understanding, multilingual support, and extended context windows (up to 128K).
The instruction-tuned Gemma-3-4B model is now available on the Clarifai Platform. Try it out for free!
👉 Gemma-3-4B-it: clarifai.com/gcp/generate/models/gemma-3-4b-it -
Clarifai
@clarifai
Create a dedicated Compute Clusters in just a few lines using Python!
Compute Clusters allow you to allocate dedicated resources to run your AI workloads more efficiently, giving you better control over performance and scaling.
With the Python SDK, you can easily create a Compute Cluster using the create_compute_cluster() method from the User class.
To create a Compute Cluster, you'll need:
✅ Your Clarifai PAT (Personal Access Token)
✅ Your Clarifai User ID
✅ A YAML config file with cluster configurations
Check out the code here: docs.clarifai.com/sdk/compute-orchestration/set-up-compute -
Clarifai
@clarifai
Clarifai at NVIDIA GTC 2025 next week! 🚀
The countdown is on! Next Week from March 17–21, join @mattzeiler, CEO & Founder of Clarifai, and our team in San Jose, CA, for NVIDIA GTC 2025.
Attending GTC? Book a meeting with the team to discover how to optimize and orchestrate your AI workloads:
👉 www.clarifai.com/events-clarifai-at-nvidia-ai-summit-2025
#nvidiagtc #computeorchestration -
Clarifai
@clarifai
🚀 HumanX is happening from today at Fontainebleau Las Vegas, and we’re excited to be part of it!
Our CEO & Founder, @mattzeiler, is attending—reach out to Matt to connect and chat about orchestrating your AI workloads and deploying any model on any compute, at any scale.
Find more details here: www.humanx.co
See you there!
#humanx -
Clarifai
@clarifai
🚀 HumanX is happening from today at Fontainebleau Las Vegas, and we’re excited to be part of it!
Our CEO & Founder, @mattzeiler, is attending—reach out to Matt to connect and chat about orchestrating your AI workloads and deploying any model on any compute, at any scale.
Find more details here: www.humanx.co
See you there!
#humanx -
Clarifai
@clarifai
Scale to Zero with Clarifai Compute Orchestration! 🚀
Scaling to zero helps you cut down infrastructure costs by automatically shutting down idle compute resources, making it perfect for workloads with variable demand.
With Clarifai's Compute Orchestration, you can configure Node Auto Scaling Range to dynamically scale resources — from handling traffic spikes to scaling down to zero when idle — ensuring you only pay for what you use.
This feature is especially useful for:
- Sporadic workloads like video or image processing
- Development and testing environments
- AI inference tasks with fluctuating traffic
👉 Learn when to scale to zero, how to set it up, and key scenarios where it makes the most impact:
www.clarifai.com/blog/how-scaling-to-zero-optimizes-ai-in… -
Clarifai
@clarifai
We have built a RAG app to chat with any GitHub repository!
Built using @clarifai, Embedchain, and @streamlit.
Features:
• Repository Loading: Load any public GitHub repository using its URL.
•Intelligent Chat:Ask questions about the code, structure, or functionality of the repository.
•Persistent Context:The app maintains context throughout the conversation for more seamless interactions.
👉 Try it out here: clarifai.com/sumanth/AI-Apps/installed_module_versions/ch…
P.S: Log in to the Platform to try out the module! -
Clarifai
@clarifai
Clarifai at HPE Tech Jam 2025 🚀
Join Joshua Wolfe, Public Sector Account Executive, and Teddy Oswari, Senior Solution Engineer, on March 24–28 in Orlando, Florida.
If you're attending, book a meeting with the team to discover how to optimize and orchestrate your AI workloads.
👉 www.clarifai.com/events/hpe-tech-jam-event-2025
#hpetechjam -
Clarifai
@clarifai
Optimize your AI workloads with cost-efficient Spot Instances! 💪
Within Compute Orchestration, you can now enable Spot Instances when configuring your nodepools — allowing you to leverage unused compute capacity at a lower price without compromising performance.
Spot Instances help you reduce costs while scaling your AI workloads efficiently.
👉 Learn how to set up Spot Instances here:
docs.clarifai.com/portal-guide/compute-orchestration/set-… -
Clarifai
@clarifai
Clarifai at the 2025 Artificial Intelligence Summit 🚀
Join Alfredo Ramos, SVP of Platform, and Joshua Wolfe, Public Sector Account Executive, at the 2025 Artificial Intelligence Summit on March 20 in McLean, VA at Jones Branch Drive.
If you're attending, book a meeting with the team to discover how to optimize and orchestrate your AI workloads:
👉 www.clarifai.com/events/clarifai-at-artificial-intelligen… -
Clarifai reposted
Matt Zeiler
@mattzeiler
Want your AI fast? The @ArrowGlobal and @clarifai partnership have got you covered! 🚀 🏁
#firestonegp -
Clarifai
@clarifai
🚀 Heading to NVIDIA GTC 2025?
Catch @mattzeiler, CEO & Founder of Clarifai and Clarifai team, from March 17–21 in San Jose, CA!
We recently announced the public preview of Compute Orchestration, and we're receiving great response from the community.
If you're attending, book a meeting with the team to learn how to optimize and orchestrate your AI workloads
👉 www.clarifai.com/events-clarifai-at-nvidia-ai-summit-2025
#gtc25 -
Clarifai
@clarifai
🚀 Catch @mattzeiler, CEO and Founder of Clarifai, at the Arrow Firestone Grand Prix of St. Petersburg in Florida from Feb 28 – Mar 2!
If you’d like to connect or sync up, feel free to reach out to Matt directly. See you there! 🏁 -
Clarifai
@clarifai
The latest edition of the AI in 5 newsletter with Clarifai is out!
Here is the summary of what we will be covering this week:
• Log and monitor every platform activity with the new Audit Logging
• Optimizing LLMs: vLLM vs LMDeploy vs SGLang
• GCP support for Compute Orchestration
• Benchmarking top vision language models (VLMs) for image classification
• Tip of the week: Access DeepSeek-R1 distilled models with multiple API clients
Read more here:
www.linkedin.com/pulse/32b-open-source-deepseek-distilled… -
Clarifai
@clarifai
Deploy AI workloads on GCP with Compute Orchestration! ⚡
Compute Orchestration now supports Google Cloud Platform (GCP). Public Preview customers can spin up dedicated compute clusters in GCP’s us-east4 region, with more regions coming soon!
👉 Learn more & sign up for the public preview here: www.clarifai.com/products/compute-orchestration -
Clarifai
@clarifai
We have Open Sourced Data Utils Library! 🚀
Data Utils library offers a suite of utilities to effortlessly handle various types of multimedia data.
You can seamlessly integrate it with the Clarifai Python SDK to unlock AI-driven solutions for both visual and textual use cases.
⮑ Find the Repo here: github.com/Clarifai/clarifai-python-datautils -
Clarifai
@clarifai
Scaling LLM inference: vLLM vs LMDeploy vs SGLang! ⚡️
Optimizing large language models is all about performance, cost, and scalability. Our latest breakdown compares these frameworks to help you choose the right one.
⮑ Read more: clarifai.com/blog/comparing-vllm-lmdeploy-and-sglang -
Clarifai
@clarifai
Deploy any model on any compute at any scale! 🚀
Clarifai’s AI Compute Orchestration enables serverless autoscaling in any environment, regardless of deployment location or hardware, allowing you to scale resources effortlessly.
Watch this short 45-second video to learn how Compute Orchestration simplifies your AI workload management. -
Clarifai
@clarifai
Here is how you can run Distilled DeepSeek R1 models for Free on the Clarifai Platform:
Try it out: clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-R1… -
Clarifai
@clarifai
Exciting news! Clarifai’s MLOps Engineer, Luv Bansal, is a 2025 Tech Innovation CUBEd Awards winner in the Breakthrough Technologists category! 🏆
Big thanks to @theCUBE for the honor! 🎉
🔗 www.thecube.net/awards -
Clarifai reposted
Matt Zeiler
@mattzeiler
Scaling LLM inference: vLLM vs LMDeploy vs SGLang!
Optimizing LLMs is all about performance, cost, and scalability.
Our latest breakdown compares these frameworks to help you choose the right one.
Read more: www.clarifai.com/blog/comparing-vllm-lmdeploy-and-sglang -
Clarifai
@clarifai
New: Audit Logging is now live on Clarifai!
Gain full visibility into platform activities with detailed logs of every action taken.
✅ See who did what
✅ Track resource changes
Monitor users, models, workflows, apps & more for better security & compliance with Audit Logging.
🔹 Use the Teams & Logs tab in the Control Center
Learn more: docs.clarifai.com/portal-guide/control-center/teams-logs/
🔹 Track audits programmatically via API
Learn more: docs.clarifai.com/api-guide/audit-log/ -
Clarifai
@clarifai
Access DeepSeek-R1-Distill-Qwen-32B via an API! 🚀
DeepSeek-R1-Distill-Qwen-32B is a 32B-parameter dense model distilled from DeepSeek-R1 based on Qwen-32B.
The model outperforms OpenAI-o1-mini across various benchmarks
Try it out or access it via an API: -
Clarifai
@clarifai
🚀 What’s the best Vision Language Model for Image Classification?
We tested GPT-4o, Qwen2-VL-7B, MiniCPM-V, and more. Which one is the best for image classification?
Check out the benchmarking results!
Read the full blog post here: www.clarifai.com/blog/best-vision-language-models-vlms-fo… -
Clarifai reposted
Matt Zeiler
@mattzeiler
🚀 Excited to announce the Public Preview of Clarifai Control Center! A unified dashboard to monitor AI ops, streamline insights, and enhance decision-making.
Press Release: www.clarifai.com/press-release/new-clarifai-control-cente… -
Clarifai
@clarifai
🚀 Clarifai 11.1 is here!
We are excited to introduce the Public Preview of Control Center: a unified dashboard for all your Clarifai activities.
Plus, we’ve added the latest DeepSeek-R1 models, GCP support for Compute Orchestration, Python SDK updates, and more!
Read more: -
Clarifai
@clarifai
🚀 We are excited to announce the Public Preview of Clarifai Control Center!
Monitor your AI operations, streamline insights, and make faster, data-driven decisions—all in one unified dashboard.
Read the Press Release here: www.clarifai.com/press-release/new-clarifai-control-cente… -
Clarifai
@clarifai
AI + Computer Vision = Faster, smarter defect detection in manufacturing! 🚀
✅ Detect defects quickly
✅ Improve efficiency
✅ Use object detection for static defects or LVMs for complex quality control
Clarifai makes AI-powered visual inspection easy!
Learn more: -
Clarifai
@clarifai
🚀 Clarifai at AI Dev World!
Patrick Lundquist, Staff Engineer at Clarifai, will be at AI Dev World to discuss Simplifying AI Development with Orchestration, covering how to:
✅ Save developer time with SDKs & CLI tools
✅ Effortlessly manage configurations & dependencies
✅ Unify AI development with a centralized control plane
📍 Santa Clara Convention Center, CA
📅 Feb 13, 1 PM
If you're attending, schedule a meeting with Patrick here: www.clarifai.com/events/clarifai-at-ai-dev-world -
Clarifai
@clarifai
vLLM vs LMDeploy vs SGLang ⚡
Looking to optimize and deploy large language models efficiently?
Our latest blog dives into vLLM, LMDeploy, and SGLang, comparing their performance, ease of use, and scalability. Whether you're a researcher or an engineer, understanding these frameworks can help you choose the best fit for your project.
What you’ll learn:
✅ Key differences between vLLM, LMDeploy, and SGLang
✅ Performance benchmarks and deployment strategies
Read the blog here:
www.clarifai.com/blog/comparing-vllm-lmdeploy-and-sglang -
Clarifai
@clarifai
🚀 Clarifai x Arrow Electronics!
We’re teaming up with @ArrowGlobal to accelerate AI adoption worldwide!
With Arrow as our official commercial distributor, businesses now have seamless access to Clarifai’s full-stack AI platform for computer vision, NLP, and AI orchestration at any scale.
Read more here: www.clarifai.com/press-release/clarifai-partners-with-arr… -
Clarifai
@clarifai
Happening Tomorrow!
Join @mattzeiler, Clarifai’s founder & CEO, for an AMA tomorrow at 12 PM ET on how to Scale and Optimize your AI workloads.
👉 Register here: www.clarifai.com/webinar-founders-ama-maximize-the-value-… -
Clarifai
@clarifai
🚀 DeepSeek's AI Revolution 🚀
Our CEO @mattzeiler shares insights on how open-source innovations and an automated feedback loop are challenging industry giants and making next-gen models more affordable.
Read more here: www.technologyreview.com/2025/01/31/1110740/how-deepseek-… -
Clarifai reposted
Matt Zeiler
@mattzeiler
Join me on Wednesday, February 5, for an AMA/webinar where I’ll share how Clarifai’s Compute Orchestration simplifies AI workload management for cost-efficient scaling and optimization.
Link to Register: www.clarifai.com/webinar-founders-ama-maximize-the-value-… -
Clarifai
@clarifai
🚀 Take control of your AI workloads and maximize the value of your investments.
Join Matt Zeiler, CEO of Clarifai, to learn how Clarifai Compute Orchestration can:
✅ Streamline AI development
✅ Cut compute costs
✅ Scale with flexibility & security
📅 Feb 5, 12 PM ET | 9 AM PT
👉 Register now: www.clarifai.com/webinar-founders-ama-maximize-the-value-…
#ai #computeorchestration #webinar -
Clarifai
@clarifai
🔥 DeepSeek models are now on Clarifai! 🔥
🚀 Try distilled DeepSeek models for free—optimized for inference-time computing to deliver smarter AI at lower costs.
Clarifai’s Compute Orchestration lets you deploy AI at scale without bottlenecks.
🔗 Access the models & read the press release here: www.clarifai.com/press-release/deepseek-now-available-on-…
#ai #deepseek #clarifai -
Clarifai
@clarifai
The latest edition of the AI in 5 newsletter with Clarifai is out!
Here is the summary of what we will be covering this week:
• DeepSeek-R1: The best open-source reasoning model
• Why Inference Optimization is key in the era of Open Source models
• Webinar: Maximize the value of your AI investments
• Tip of the week: Leverage Unused Compute with Spot Instances
Read here: www.linkedin.com/pulse/introducing-deepseek-r1-best-open-…
#deepseekr1 #computeorchestration -
Clarifai
@clarifai
⏳ 1 week left!
Join Matt Zeiler on Feb 5 at 12 PM ET for an AMA on Compute Orchestration:
🚀 Optimize AI workloads
💡 Scale effortlessly
💰 Save costs
Whether you're a developer, AI practitioner, Infra Ops engineer, or AI enthusiast, this session will provide you with the insights and tools to streamline, optimize, and scale your AI workloads effectively.
Register here: www.clarifai.com/webinar-founders-ama-maximize-the-value-… -
Clarifai reposted
Matt Zeiler
@mattzeiler
DeepSeek-R1 is a leap for open-source AI, driving innovation.
But the real challenge? Inference optimization. As AI grows, scalable, efficient solutions are key.
Open-source has huge potential, we’re here to unlock it.
www.clarifai.com/blog/optimized-inference-in-the-age-of-o… -
Clarifai
@clarifai
We are excited to share that the DeepSeek-R1 Distilled Models are Now on Clarifai and you can try them out for Free! 🚀
DeepSeek-R1 sets a new standard for open-source reasoning, rivaling OpenAI-o1 in performance while being faster and more cost-effective. These distilled models, derived from DeepSeek-R1 based on Qwen, deliver top-tier results in math, coding, and reasoning tasks, optimized for efficiency.
Key Highlights: 🔥
• DeepSeek-R1-Distill-Qwen-32B outperforms OpenAI-o1-mini across various benchmarks, achieving new state-of-the-art results for dense models.
• These models are built for speed and cost-efficiency.
We have made them free for the community users for a limited time. Try them out!
1. DeepSeek-R1-Distill-Qwen-1.5B: clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-R1…
2. DeepSeek-R1-Distill-Qwen-7B: clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-R1…
3. DeepSeek-R1-Distill-Qwen-14B: clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-R1…
4. DeepSeek-R1-Distill-Qwen-32B: clarifai.com/deepseek-ai/deepseek-chat/models/DeepSeek-R1…
#deepseekr1 -
Clarifai
@clarifai
DeepSeek AI recently released DeepSeek-R1—the most advanced open-source reasoning model, rivaling OpenAI's o1!
But here's the real challenge: inference optimization. As more companies adopt these models, demand for scalable, efficient inference will skyrocket.
Learn why it’s key in the age of open-source innovation
www.clarifai.com/blog/optimized-inference-in-the-age-of-o…
#ai #opensource #inference #deepseekr1 -
Clarifai
@clarifai
AI is revolutionizing businesses, but it’s complex and costly. Compute Orchestration is the solution.
Join @mattzeiler, CEO of Clarifai, to explore how it can:
✅ Optimize AI workloads
✅ Cut costs by up to 90%
✅ Enable real-time, scalable deployments
📅 February 5 at 12 PM ET | 9 AM PT
Reserve your spot: www.clarifai.com/webinar-founders-ama-maximize-the-value-… -
Clarifai
@clarifai
Clarifai Compute Orchestration brings serverless autoscaling to any environment.
Join @mattzeiler, Founder & CEO, on Feb 5 at 12 PM ET | 9 AM PT to learn how to:
✅ Deploy AI models on any infrastructure
✅ Use cost-efficient nodepools
✅ Customize for performance & security
👉 RSVP here: www.clarifai.com/webinar-founders-ama-maximize-the-value-…
#computeorchestration -
Clarifai
@clarifai
Exclusive AMA with Founder & CEO @mattzeiler
Join us on February 5 at 12 PM ET | 9 AM PT to explore how Clarifai’s new Compute Orchestration transforms AI deployment.
Learn how to streamline AI workloads, leverage real-time edge processing, and maximize cost efficiency.
👉 Register now: www.clarifai.com/webinar-founders-ama-maximize-the-value-…
#ai #webinar #computeorchestration -
Clarifai
@clarifai
New Release: Clarifai 11.0 is Out Now! ⚡️
We’re introducing Control Center, a unified dashboard that serves as a single pane of glass to monitor everything happening within your account on the Clarifai Platform.
The Control Center acts as a single source of truth for multiple data dimensions, enabling you to make informed decisions with insights from various sources.
In this release, we’ve also introduced the Financial Manager role for organizations, made significant updates to the Python SDK, and included various other improvements and bug fixes.
Read more in the release blog here: clarifai.com/blog/clarifai-11.0-clarifais-control-center
#controlcenter -
Clarifai
@clarifai
We are thrilled to announce that Clarifai’s own ML Ops Engineer Luv Bansal has been named a finalist in the Breakthrough Technologists category at the Tech Innovation CUBEd Awards! 🏆
www.thecube.net/awards -
Clarifai
@clarifai
🚀 Exciting news!
Clarifai’s Compute Orchestration Capability is featured in @Techstrongai for its game-changing ability to cut AI compute costs by up to 90% while optimizing performance. 💡📈
Read more: techstrong.ai/articles/clarifai-previews-compute-orchestr…
#ai #machinelearning #costoptimization -
Clarifai
@clarifai
Managing AI projects across teams can be challenging, involving different tools, environments, and resource needs.
On top of that, tracking performance, controlling costs, and optimizing deployments add even more complexity!
That’s where Clarifai's Control Center comes in.
With this single pane of glass, you can:
* Monitor models and compute in real time.
* Manage AI resources, performance, and costs effortlessly.
Ensure your AI adoption is governed efficiently in one place.
👉 Read more about Control Center here: docs.clarifai.com/portal-guide/control-center/ -
Clarifai
@clarifai
In today’s fast-paced AI landscape, teams are often struggling with two major challenges!
• Inconsistent AI tooling across teams: Multiple teams working on diverse AI projects without a reference architecture leads to duplicate and inconsistent practices.
• Inefficient AI development lifecycles: Teams spend 75% of their time building AI infrastructure, facing delays in resource access, and lacking overall visibility.
Clarifai addresses both by offering a full-stack AI platform that unifies tools and practices, while Compute Orchestration enables users to streamline model deployments and optimize compute resources across any environment, whether SaaS, VPC, on-prem, or air-gapped.
Looking to optimize your AI workflow?
👉 Read more here: clarifai.com/products/compute-orchestration -
Clarifai
@clarifai
Marketers, ready to level up in 2025?
Watch our on-demand webinar "AI for Marketers: Content Generation and Personalization with Visual Search!"
Learn how to use #ai-powered #contentpersonalization to drive more conversions. www.clarifai.com/webinar-ai-for-marketers-content-generat…
#aiformarketing #marketing -
Clarifai
@clarifai
Happy New Year from all of us at Clarifai! ✨
2024 was a blast, but we’re excited for what’s coming next. 🚀
Big things are brewing at Clarifai, and we can’t wait to share all the AI magic with you.
Thank you for your love and support. 💙
Here’s to making 2025 unforgettable. 🥂 -
Clarifai
@clarifai
76% consumers feel let down if they don’t get personalized shopping suggestions. 😲
Surely, you don't want to be the one to disappoint your customers and lose sales‼️
Learn how you can offer a personalized digital experience with #visualsearch and #ai. 👇
www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
🥁 Drum roll please.
Clarifai launches AI Sprint Workshops to help organizations rapidly identify & deploy AI solutions—even if they lack in-house AI experience. Go from prototype to production in days! Learn more: www.clarifai.com/press-release/clarifai-launches-ai-sprin…
#aiinnovation #aiworkshops #adoptai -
Clarifai
@clarifai
Capturing key moments with industry leaders at NeurIPS 2024!
Clarifai's CEO @mattzeiler and Sr. Director of Research Aaron Reite had an insightful discussion with the Google team.
Stay tuned for more updates from NeurIPS!
If you are NeurIPS, book some time to chat with some of our team onsite: www.clarifai.com/events-clarifai-at-neurips-2024
#neurips2024 -
Clarifai
@clarifai
Clarifai is at @NeurIPSConf 2024! Join us in Vancouver Convention Centre from December 10-15 🚀
Creating AI prototypes can seem easy at first. But, building a production-grade AI application that can support an enterprise at scale can take years.
Meet our experts at the NeurIPS 2024 to learn more about Clarifai and how we can help you build and deploy production-grade AI in minutes!
Book time here to chat to some of our team onsite: www.clarifai.com/events-clarifai-at-neurips-2024
#neurips2024 #ai #machinelearning -
Clarifai
@clarifai
Announcing a strategic engagement with @GettyImages, integrating their #genai features into the #clarifai platform. 👏
Powered by @NVIDIAAI and Edify, and trained on Getty Images’ library, Clarifai users can access responsibly developed #genai visuals. www.clarifai.com/press-release/clarifai-and-getty-images-… -
Clarifai
@clarifai
Enhance customer shopping experience, boost product discoverability, and drive sales with one tool. 💰
With #computervision and #similaritysearch, you can enable your customers to find items that are similar to their taste.
Watch our webinar to learn how. www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
Clarifai received the top score in the current offering category and is recognized as a Leader in The #forrester Wave™: #computervision Tools, Q1 2024. 🎉
Check out the full report to learn more about our offerings and how #clarifai stacks up. ➡️ www.clarifai.com/report/the-forrester-wave-computer-visio… -
Clarifai
@clarifai
Unlock the secret to accurate product search recommendations. 🔓
Watch our webinar to learn how you can use #ai and #visualsearch to tag, classify and search assets that help you deliver a personalized web experience and generate more conversions! www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
With #awsreinvent this week, we know a lot of our customers are thinking about how to get the most out of their compute investments! 🚀
Clarifai’s new compute orchestration capabilities are a game-changer for enterprises seeking to better leverage both their cloud and hardware commitments for AI, thanks to a unified control plane that connects to and manages multiple AI compute clusters.
Leverage your cloud commitments or hardware investments while benefiting from serverless autoscaling, whether it's SaaS, VPC, on-prem, or air-gapped.
Read the press release here: www.clarifai.com/press-release/clarifai-launches-first-ve…
📅 Or book time to chat with some of our team onsite in Las Vegas: www.clarifai.com/events-clarifai-at-aws-reinvent-2024
#ai #awsreinvent #reinvent2024 -
Clarifai
@clarifai
🎉 Excited to announce the public preview of Clarifai’s new compute orchestration capabilities! 🚀
With this game-changer, enterprises can:
✅ Maximize compute investments
✅ Leverage cloud & hardware commitments for AI
✅ Centrally manage costs & performance
Eliminate vendor lock-in with the only platform for AI across any environment.
Learn more: www.clarifai.com/press-release/clarifai-launches-first-ve… -
Clarifai
@clarifai
Just 2 days to go until AWS re:Invent! 🚀
We are counting down to AWS re:Invent in Las Vegas from December 2-6, 2024! Our team is excited to share how Clarifai’s AI solutions can transform your business.
🤝 Book your meeting here: www.clarifai.com/events-clarifai-at-aws-reinvent-2024
See you in Vegas! -
Clarifai
@clarifai
All set to hit AWS re:Invent? 🔥
Join us to see how Clarifai is leading the way in AI innovation! Our team is eager to meet you and showcase our latest solutions that are transforming industries.
📅 Date: December 2-6, 2024
📍Location: Las Vegas, NV
🤝 Want to meet? Schedule a one-on-one session with our team!
Set up your meeting now: www.clarifai.com/events-clarifai-at-aws-reinvent-2024
We can't wait to connect with you! -
Clarifai
@clarifai
AI is no longer a nice-to-have; it’s a must-have. Where is your organization on the AI journey?
With AWS re:Invent just around the corner, now is the time to focus on the topics that matter most to you.
Join us to explore what we’re building and experience impressive demos from our experts.
Check out the page and register here: www.clarifai.com/events-clarifai-at-aws-reinvent-2024
Our team will be in touch with you soon! 🚀
#awsreinvent -
Clarifai
@clarifai
🚀 Boost Your Online Sales with AI This Black Friday! 🛍️
AI is revolutionizing online retail—69% of retailers see increased revenue after adopting it. Here’s how you can leverage Computer Vision to stay ahead this holiday season:
1. Enhance Data Quality: Use AI to refine product metadata for better search results and higher conversions.
2. Visual Search: Let customers snap and find products they love instantly without relying on keywords.
3. Personalized Shopping: AI-driven recommendations boost repeat purchases and customer satisfaction.
It’s not too late to future-proof your retail strategy with AI.
Read more in the blog post here: www.mytotalretail.com/article/3-ways-online-retailers-can… -
Clarifai
@clarifai
Today, we celebrate 11 years of Clarifai! 🎉
Over the past decade, we’ve pioneered AI innovations that have transformed industries, empowered organizations, and redefined what’s possible. From launching the groundbreaking V1 API for image recognition to introducing tools ahead of their time: CUDA kernel optimizations before they were industry-standard, Kubernetes GPU operators preceding even NVIDIA, and platforms that outpaced hyperscalers; we’ve consistently pushed boundaries. Today, billions of AI operations flow through our platform, powering enterprises, governments, and innovators worldwide.
Our journey is defined not just by technology, but by the incredible people behind it. Our team’s passion for innovation drives us to create the most accessible and powerful AI platform, enabling seamless collaboration and empowering entire teams to build world-class AI solutions.
As we celebrate this milestone, we look to the future with exciting innovation that helps apply AI to solve humanity’s greatest challenges.
To everyone who’s been part of this journey: THANK YOU. Here’s to 11 years of innovation and the infinite possibilities ahead! -
Clarifai
@clarifai
Get Ready for AWS re:Invent 2024 with Clarifai! 🚀
We’re thrilled to partner with Carahsoft to showcase the latest in AI innovation! Join us in Las Vegas, NV, from December 2–10, and discover how Clarifai’s advanced solutions can transform your business.
🌟 Why Attend?
• Build and deploy production-ready AI in minutes.
• Connect with our AI experts and explore cutting-edge solutions.
• Network with industry leaders and innovators.
⭐ Meet the Experts
Our team Matthew Zeiler, Douglas Shapiro, Alfredo Ramos, Jonathan Padgett, Patrick Connell, and Pratik C. will be there to demo powerful AI solutions and answer your questions. Don’t miss this chance to learn from the best.
Learn more here: www.clarifai.com/events-clarifai-at-aws-reinvent-2024 -
Clarifai
@clarifai
Tune in to our CEO @mattzeiler conversation with the @StackOverflow team, where they discuss how transformer architecture has replaced convolutional neural networks in AI applications, the impact of AI regulation, and the growing importance of synthetic data!
Stack Overflow
@StackOverflow
🎙️The podcast team sits down with @mattzeiler, founder and CEO of @clarifai, to chat about how the transformer architecture supplanted convolutional neural networks in AI applications, the implications of regulating AI, and the value of synthetic data.
stackoverflow.blog/2024/11/12/a-student-of-geoff-hinton-y… -
Clarifai
@clarifai
Clarifai joins the BAIR Open Research Commons! 📷We're proud to work alongside @berkeley_ai to advance cutting-edge AI, from multimodal learning to scalable, real-world solutions! 🎉
Here’s to an impactful collaboration and the future of applied AI!
Read here: -
Clarifai
@clarifai
Ready for NeurIPS 2024! 🚀
We’re excited to announce our participation at NeurIPS 2024! Join us in New Orleans from December 10-16, 2024, to explore the forefront of AI and machine learning innovation.
Discover how Clarifai is pushing the boundaries of what's possible with AI. Meet our team of AI visionaries Matthew Zeiler, Mark Lowell, Aaron Reite, Peter Rizzi and Christopher Mullarkey.
Stop by our booth for live demos and to see firsthand how Clarifai's AI tools can optimize your operations, enhance decision-making, and drive innovation.
RSVP here: www.clarifai.com/events-clarifai-at-neurips-2024 -
Clarifai
@clarifai
Tomorrow is the day - Come meet Mark Lowell, our Staff Research Scientist, who will be presenting at the ATRWG event hosted by the DEVCOM C5ISR Center and DEVCOM Aviation and Missile Center (AvMC) at the Science and Engineering Laboratory! 🚀
Date: November 13, 2024
Location: ATRWG Event www.vdl.afrl.af.mil/atrwg/
Mark will be showcasing how Clarifai’s cutting-edge AI and machine learning solutions are revolutionizing public sector operations. Don’t miss out on this exciting opportunity to learn how Clarifai is at the forefront of AI advancements for government operations. Be part of the future with us!
Stay tuned for updates and insights from the event. -
Clarifai
@clarifai
We are excited to announce that our Staff Research Scientist, Mark Lowell, will be presenting at the ATRWG event hosted by the DEVCOM C5ISR Center and DEVCOM Aviation and Missile Center (AvMC) at the Science and Engineering Laboratory!
Event Date: 13/11/2024
Location: ATRWG Event www.vdl.afrl.af.mil/atrwg/
Mark will discuss Clarifai's role in AI and machine learning advancements for the public sector, focusing on solutions that enhance government operations and drive innovation. -
Clarifai
@clarifai
✨ Good AI shouldn’t just work in the lab. It should work where the magic happens—on your production floor.
Our whitepaper shows you how to effortlessly operationalize your AI into production at scale.
Download it now.
www.clarifai.com/report/an-operating-model-for-orchestrat… -
Clarifai
@clarifai
Say goodbye to human error and hello to automated #defectdetection that’s fast, scalable, and keeps your employees safe.
📥 Download the #visualinspection Solution Brief and upgrade your #predictivemaintenance today: www.clarifai.com/visual-inspection-solutions-by-clarifai?…
#ai #computervision -
Clarifai
@clarifai
IBM introduced the new Granite 3.0 models last week! 🎉
These models are designed for enterprise AI, offering robust performance for tasks such as RAG, classification, summarization, entity extraction, and tool utilization.
Compact and versatile, they are built to be fine-tuned with enterprise data and seamlessly integrated into diverse business environments and workflows.
The models are now available on the Clarifai Platform. Try them out and access with an API: 👇
Granite-3.0-8B-Instruct: clarifai.com/ibm/granite/models/granite-3_0-8b-instruct
Granite-3.0-2B-Instruct: clarifai.com/ibm/granite/models/granite-3_0-2b-instruct -
Clarifai
@clarifai
🤔 Your AI prototype is great... but can it work in production?
Our whitepaper explains how to easily operationalize your AI at scale, eliminating broken pipelines and integration issues.
📄 Get the whitepaper and turn your AI into production: www.clarifai.com/report/an-operating-model-for-orchestrat… -
Clarifai
@clarifai
Exciting News! 🚀
Our Senior Director of Research, Aaron Reite, is presenting today at the Military Sensing Symposia (MSS) Conference, sharing his expertise on two crucial topics: "Effective RAG for Air-Gapped Systems" and "Leveraging LLMs for Intel Analysis."
Join us to explore how Clarifai's innovative AI and machine learning solutions are transforming secure data environments and revolutionizing intelligence analysis. This is a fantastic opportunity to see how we're advancing real-world applications for high-security needs!
Learn more about the event here: www.mssconferences.org/public/meetings/conferenceDetail.a…
#aiinnovation #machinelearning #mssconferences #securedata #intelanalysis -
Clarifai
@clarifai
Manual inspections are cool... if it’s 1924.
🚀 Fast forward to #ai-powered visual inspections—detect defects faster, reduce downtime, and keep your processes moving smoothly.
📥 Download the solution brief to learn how to get started: www.clarifai.com/visual-inspection-solutions-by-clarifai?… -
Clarifai
@clarifai
🎉 We’re excited to announce that our Senior Director of Research, Aaron Reite, will present at the Military Sensing Symposia (MSS) Conference on October 30, 2024, sharing his expertise on two critical topics: "Effective RAG for Air-Gapped Systems" and "Leveraging LLMs for Intel Analysis."
With Clarifai's cutting-edge approach to AI and machine learning, Aaron will explore solutions for secure data environments and innovative applications of large language models in intelligence analysis.
Don't miss this opportunity to learn how Clarifai is transforming real-world applications for high-security needs!
Learn more about the event here: www.mssconferences.org/public/meetings/conferenceDetail.a… -
Clarifai
@clarifai
⚠️ Prototypes are easy. Production-ready AI? Not so much.
Our whitepaper shows how to operationalize and orchestrate AI so your project doesn’t get stuck in “prototype nightmare.”
Ready to go from patched up demo to AI in action? Download our whitepaper. www.clarifai.com/report/an-operating-model-for-orchestrat… -
Clarifai
@clarifai
With robust prebuilt app templates and quick-to-implement #cv workflows, choose Clarifai for innovative and scalable #ai.
Read the Forrester Wave™: Computer Vision Tools, Q1 2024 report to explore the full breadth of Clarifai’s CV offerings. www.clarifai.com/report/the-forrester-wave-computer-visio… -
Clarifai
@clarifai
Access Pixtral-12B, a multimodal language model from Mistral AI, now available on the Clarifai Platform!
The model can process both text and visual inputs, including reasoning with charts, figures, and natural scenes.
It also offers a long context window of 128K tokens, allowing efficient management of multiple images and large volumes of text.
Try out the model on the Clarifai platform or access it via API: clarifai.com/mistralai/completion/models/pixtral-12b -
Clarifai
@clarifai
👀 Still squinting at defects? Time to let AI take a closer look.
Swap manual guesswork for automated precision with our Visual Inspection Solution Brief—it's like eagle eyes for your production line.
📥 Download now and let your quality control soar: www.clarifai.com/visual-inspection-solutions-by-clarifai?… -
Clarifai
@clarifai
🎟️ Will you be at AI/ML & Analytics Technology Exchange & Expo in Orlando tomorrow?
Let’s chat!
Learn how we’re helping enterprises move from prototypes to production-grade AI in minutes. 💪🚀
Book a meeting with our on-site experts - www.clarifai.com/events-clarifai-at-ai-ml-technology-exch…
#ai #ml #aimlexpo -
Clarifai
@clarifai
Introducing the new Clarifai Control Center, the unified dashboard, a single pane of glass to monitor everything happening within your account on the platform!
Control Center helps streamline the management of your Clarifai operations by consolidating all activities into a single interface, minimizing the need to switch between different tools or windows.
Check out the quick tutorial below to learn more about Control Center:
youtu.be/9-yNRZsqWdM -
Clarifai
@clarifai
🔍 Relying on humans to spot every flaw? Good luck.
With #computervision and #visualinspection, automate inspections, reduce downtime, and boost efficiency— all while keeping your team safe from hazards.
Download our solution brief to get started - www.clarifai.com/visual-inspection-solutions-by-clarifai?… -
Clarifai
@clarifai
🚀 Ready to turn your AI prototypes into production-grade solutions? Meet us at the AI/ML & Analytics Technology Exchange & Expo in Orlando! 🏙️
📅 Oct 22
✅ Connect with our experts and learn how to build AI that scales—fast!
Book an on site meeting - www.clarifai.com/events-clarifai-at-ai-ml-technology-exch… -
Clarifai
@clarifai
🎯Clarifai is the “Swiss Army knife” for computer vision, according to our customers!
Transform your #ai prototypes into a production-ready infrastructure.
Read more about Clarifai’s CV offerings and how they stack up against other top players. www.clarifai.com/report/the-forrester-wave-computer-visio… -
Clarifai
@clarifai
We're thrilled to announce that Clarifai has joined the @HPE Technology Partner Program as a Silver Partner! 🎉
Together, we're making AI + computer vision solutions more accessible, faster to deploy and secure at scale.
Read more: www.clarifai.com/press-release/clarifai-joins-hewlett-pac… -
Clarifai reposted
Matt Zeiler
@mattzeiler
Thrilled to have spoken at the Industrial Future Summit 2024 in Stockholm yesterday, sharing insights on optimizing operational efficiency in industrial applications using Clarifai.
The energy at the event was incredible! -
Clarifai
@clarifai
Hello from Stockholm!
Our CEO, @mattzeiler is attending the Stockholm Industrial Summit tomorrow, October 15.
Want to improve operational efficiency of your industrial project? Book a meeting with Matt. www.clarifai.com/events-book-a-meeting-with-matt-zeiler?h…
#ai #visualinspection #defectdetection -
Clarifai
@clarifai
Clarifai is going to the Stockholm Industrial Summit on October 15.
Book a meeting to chat with our CEO, @mattzeiler about how you can improve operational efficiency of your industrial applications. www.clarifai.com/events-book-a-meeting-with-matt-zeiler?h…
#ai #visualinspection #defectdetection -
Clarifai reposted
Matt Zeiler
@mattzeiler
Excited to be at the @AMD Advancing AI 2024 event today!
Proud to see @clarifai featured as one of the leading generative AI platforms powering AI workloads on their cutting-edge hardware. -
Clarifai
@clarifai
Clarifai is featured at today's @AMD Advancing AI 2024 event as one of the leading generative AI platforms powering AI workloads on their cutting-edge hardware.
Matt Zeiler
@mattzeiler
Excited to be at the @AMD Advancing AI 2024 event today!
Proud to see @clarifai featured as one of the leading generative AI platforms powering AI workloads on their cutting-edge hardware. -
Clarifai
@clarifai
New Release: Clarifai 10.9 is Out Now! ⚡️
We’re excited to introduce the new Clarifai Control Center, a unified dashboard that provides a comprehensive view of everything happening within your account on the platform.
The Control Center serves as a single source of truth for multiple data dimensions, empowering you to make informed decisions based on insights from various sources.
In addition to the Control Center, this release also includes new models, and a private preview of our upcoming compute orchestration, along with other improvements and bug fixes.
Read more in the release blog here:
www.clarifai.com/blog/clarifai-10.9-control-center -
Clarifai
@clarifai
Learn how Clarifai stacked up in @Forrester's latest assessment of enterprise-level #computervision tools. Read the #forrester Wave™: Computer Vision Tools, Q1 2024 report to find the best CV platform for your needs. www.clarifai.com/report/the-forrester-wave-computer-visio…
#visualinspection #contentmoderation -
Clarifai
@clarifai
We're excited to announce the partnership between Clarifai and Crimson Phoenix! 🤝
This alliance aims to deliver cutting-edge AI-enabled data labeling and CV capabilities for unstructured data in the Intelligence and Defense communities.
Press release www.clarifai.com/press-release/clarifai-and-crimson-phoen… -
Clarifai
@clarifai
Curious about geospatial data and AI foundation model research? Join Aaron Riete, our Senior Director of Research, as he leads an insightful panel on Geospatial AI Insights today at the NVIDIA AI Summit. Don't miss out! www.nvidia.com/en-us/events/ai-summit/session-catalog/?ta…
-
Clarifai reposted
Sumanth
@Sumanth_077
Build a LLM app with RAG to chat with PDF files using the latest open-source Llama 3.2 model in Python:
-
Clarifai
@clarifai
We're going to Washington D.C.!
Our team will be attending the @NVIDIAAI Summit from October 7-9.
Book a meeting with our on-site experts to find out how to convert your AI possibilities to reality.
www.clarifai.com/events-clarifai-at-nvidia-ai-summit-2024… -
Clarifai
@clarifai
Are you going to the @nvidia AI Summit in Washington D.C.?
Aaron Reite, our Senior Director of Research will be participating in a panel on Geospatial AI Insights on October 9. Don't miss it. www.nvidia.com/en-us/events/ai-summit/session-catalog/?ta… -
Clarifai
@clarifai
Have you tried the latest Llama 3.2 Vision model yet? 👀
Meta AI has announced the release of Llama 3.2, introducing the first multimodal models in the series.
The 11B and 90B parameter multimodal models are designed for visual reasoning, image captioning, and VQA tasks. They can now process and understand both text and images.
The Llama 3.2 11B Vision Instruct model is now available on the Clarifai platform!
Try out the model, and easily access it via the API here: clarifai.com/meta/Llama-3/models/llama-3_2-11b-vision-ins… -
Clarifai
@clarifai
Personalized marketing strategies can help you drive more conversions. Wondering how to get started with building a solid framework for content personalization and organization? Watch our webinar on demand. www.clarifai.com/webinar-ai-for-marketers-content-generat…
#aiinmarketing #contentcreation #leadgeneration -
Clarifai
@clarifai
Clarifai, a Leader in The @forrester Wave™: Computer Vision Tools, Q1 2024, is recognized with the highest scores possible in the Vision, Innovation, Partner ecosystem, and Roadmap criteria.
Check out Clarifai’s standings in the #cv industry. www.clarifai.com/report/the-forrester-wave-computer-visio… -
Clarifai
@clarifai
Meta released Llama 3.2, featuring small and medium-sized multimodal LLMs (11B and 90B) as well as lightweight text-only models (1B and 3B) designed for mobile and edge devices! 🚀
Key Highlights:
- Vision Use Cases: The multimodal Llama 3.2 models support image reasoning tasks, such as document-level understanding (including charts and graphs), image captioning, and more.
- Performance: The 3B model outperforms the Gemma 2 (2.6B) and Phi 3.5-mini models in tasks such as instruction-following, summarization, and tool use, while the 1B model is competitive with Gemma.
- SLMs Capabilities: With a 128k-token context window, the lightweight 1B and 3B models excel at multilingual text generation and tool-calling tasks. They are ideal for building personalized, on-device agentic applications where data remains on the device.
The Llama 3.2 11B Vision Instruct and 3B Instruct models are now available on the Clarifai platform. 🎉
Try them out and access via API!
Llama 3.2 11B Vision Instruct: clarifai.com/meta/Llama-3/models/llama-3_2-11b-vision-ins…
Llama 3.2 3B Instruct: clarifai.com/meta/Llama-3/models/llama-3_2-3b-instruct -
Clarifai
@clarifai
✨ That’s a wrap! Thank you to everyone who joined our webinar on AI for Marketers!
Missed it? Don’t worry—you can still catch the highlights and key takeaways.
📽️ Watch the replay: www.clarifai.com/webinar-ai-for-marketers-content-generat…
#ai #marketingstrategy -
Clarifai
@clarifai
Ready to transform your content personalization strategy with AI?
Join us at 1 PM ET for "AI for Marketers: Content Generation and Personalization with Visual Search" and discover the game-changing power of AI.
🎯 It’s not too late—register now
👉 www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
📊 Brands using AI-powered product recommendations see a 5.5x improved conversion rate! Learn how you can do this too.
👉 Register for our webinar today: www.clarifai.com/webinar-ai-for-marketers-content-generat…
#marketingstrategy #aiformarketers #personalizedmarketing -
Clarifai reposted
Clarifai
@clarifai
Fine-Tuning Llama 3.1 Made Easy!
You can now easily fine-tune Llama 3.1 using the new training template in the Platform UI.
Whether you need extended context, enhanced instruction-following, or advanced capabilities for tasks like text generation and classification, it's all achievable.
Check out this complete tutorial on fine-tuning the Llama 3.1 8B Instruct model with your data using the Clarifai Platform:
youtu.be/J2N4AbXlWZM -
Clarifai
@clarifai
Fine-Tuning Llama 3.1 Made Easy!
You can now easily fine-tune Llama 3.1 using the new training template in the Platform UI.
Whether you need extended context, enhanced instruction-following, or advanced capabilities for tasks like text generation and classification, it's all achievable.
Check out this complete tutorial on fine-tuning the Llama 3.1 8B Instruct model with your data using the Clarifai Platform:
youtu.be/J2N4AbXlWZM -
Clarifai
@clarifai
How can AI streamline your #digitalassetmanagement?
AI can label, classify, and categorize content, automate labeling, and identify and remove duplicates.
Join our webinar to learn how to use #ai for an effective #marketing experience. www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
🔎 Why #ai is the future of #marketing?
The next-gen marketer’s best friend? AI. It’s not just hype; it’s your new growth engine.
Join our webinar on #contentgeneration and personalization with #visualsearch.
📅 Sept 25 | Register now www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
Looking for the ideal platform for your #computervision needs? 🤔
Read the @forrester Wave™ assessment of the top #cv tools vendors and discover how Clarifai can be a natural solution for public and private sector users across industries. 👉 www.clarifai.com/report/the-forrester-wave-computer-visio… -
Clarifai
@clarifai
71% consumers expect personalized product recommendations and 76% feel let down when they don’t get them.
With AI, you can tag, organize, and search your content to deliver accurate product recommendations and boost conversions.
Register to learn more. www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
OpenAI has released its new model "o1" which is capable of thinking before it answers.
This model can reason through complex tasks and solve more challenging problems than previous models in science, coding, and math.
Here are some key capabilities of the model:
- Unlike its predecessors, which prioritized pre-training, o1 invests in prolonged inference, sharpening its logical abilities.
- It’s smarter than most humans in competitive programming (top 89%) and math competitions (83% accuracy in IMO qualifiers).
- It can solve PhD-level problems in physics, biology, and chemistry.
- There are two versions: o1-preview, designed for refining use cases, and o1-mini, a smaller model that is 80% cheaper than o1-preview and delivers almost the same performance for STEM challenges.
Try out o1-preview and o1-mini on the Clarifai Platform, and access them via the API now! 🎉
o1-preview: clarifai.com/openai/chat-completion/models/o1-preview
o1-mini: clarifai.com/openai/chat-completion/models/o1-mini -
Clarifai
@clarifai
Did you know marketers spend nearly a month per year looking for content? 🤯
With #ai, you can simplify content tagging, and organizing saving time and boosting conversions.
Join us on Sept 25 for a webinar on #contentpersonalization with #visualsearch!
🎟️ www.clarifai.com/webinar-ai-for-marketers-content-generat… -
Clarifai
@clarifai
Personalized experiences are the new norm!
Curious how to deliver them? Join us for a webinar on Sept 25 to learn how visual search and AI can power your personalized content strategy.
Register now - www.clarifai.com/webinar-ai-for-marketers-content-generat…
#marketing #ai #leadgeneration #contentpersonalization -
Clarifai
@clarifai
New Release: Clarifai 10.8 is Out Now! ⚡️
We’re excited to introduce a new way to add relationships between concepts on the platform, allowing you to define hierarchical and semantic links using the Python SDK.
By mapping concepts as synonyms, hyponyms, or hypernyms, you can create context-aware models that deliver more accurate results. This is especially useful for managing data, building search engines, and organizing large datasets, ultimately improving the effectiveness of AI applications.
In addition to Advanced Concept Management, this release includes new features, models, other improvements, and bug fixes.
Read more in the release blog here:
www.clarifai.com/blog/clarifai-10.8-supercharge-ai-models… -
Clarifai
@clarifai
🚀 The future of marketing is here!
Join our webinar "AI for Marketers: Content Generation and Personalization with Visual Search" to discover how AI can transform your campaigns.
📅 Save the date: September 25, 1 PM ET
📍 Register now: www.clarifai.com/webinar-ai-for-marketers-content-generat…
#marketing #ai -
Clarifai
@clarifai
🚀 Marketers, ready to level up?
Join us for a webinar on "AI for Marketers: Content Generation and Personalization with Visual Search!"
Learn how to create personalized content that converts, powered by AI!
Register now! www.clarifai.com/webinar-ai-for-marketers-content-generat…
#marketing #ai #visualsearch -
Clarifai
@clarifai
Leverage AI to Assist and Optimize the Process of Annotating Data! 🚀
You can request predictions from any model or workflow available to you for a particular input, and then review, correct, or validate them before converting them into annotations.
AI-Assist provides several key benefits, including:
• Significantly accelerating the labelling process, reducing the time required to create labelled datasets.
• Reducing labor costs associated with manual labelling through automation.
• Providing consistent labelling with AI models, minimizing human errors.
• Allowing efficient and scalable handling of large datasets or rapidly changing data.
Check out the guide on AI-assisted labelling in this video: youtu.be/x4lC5sz-Oqs -
Clarifai
@clarifai
Fine-Tuning Llama 3.1 Made Easy!
You can now easily fine-tune Llama 3.1 using the new training template in the Platform UI.
Whether you need extended context, enhanced instruction-following, or advanced capabilities for tasks like text generation and classification, it's all achievable.
Check out this complete tutorial on fine-tunung the Llama 3.1 8B Instruct model with your data using the Clarifai Platform:
youtu.be/J2N4AbXlWZM -
Clarifai
@clarifai
Multimodal models can process text and images as inputs, and in some cases, they also handle other modalities like video and speech.
The blog below provides an overview of ten key multimodal datasets and benchmarks that can be used to evaluate the performance of such models, particularly those focused on Visual Question Answering (VQA).
Examples include TextVQA, DocVQA, OCRBench, MathVista, and more.
Read here: www.clarifai.com/blog/the-landscape-of-multimodal-evaluat… -
Clarifai
@clarifai
LLM-powered applications can be vulnerable to prompt attacks, where malicious prompts are designed to manipulate the model's behavior against the developer's intentions.
There are two types of prompt attacks:
Prompt Injections: These are inputs that exploit the inclusion of untrusted data from third parties or users into a model's context, causing the model to follow unintended instructions.
Jailbreaks: These are malicious instructions intended to override a model's built-in safety and security features.
Prompt-Guard-86M is a multilingual classifier model from Meta designed to detect and prevent these prompt injections and jailbreak attacks in LLM-powered applications.
The model is now available on the Clarifai Platform. We have some pre-built examples to help you get started:
Try it out here: clarifai.com/meta/text-classifier/models/prompt-guard-86m -
Clarifai
@clarifai
Companies that deliver a personalized digital experiences win brand loyalty.
Watch our on-demand webinar "Digital Asset Management and Personalization with AI" to learn how to generate conversions using a highly personalized content strategy. www.clarifai.com/events/webinar-digital-asset-management-…
#dam #ai -
Clarifai
@clarifai
What is RAFT? Combining RAG with Fine-tuning
RAG adds extra knowledge from outside sources to the prompts, and fine-tuning gives the model more data to learn from.
Each method has pros and cons, and the choice between them often depends on the project's needs.
RAFT combines the benefits of RAG and fine-tuning by improving the model's understanding and use of domain-specific knowledge while maintaining accuracy. This ensures that the LLM generates more accurate and contextually relevant answers.
Learn more about RAFT, its performance, and results with the Llama 3.1 8B model in this blog here: www.clarifai.com/blog/supercharge-your-llm-via-retrieval-… -
Clarifai
@clarifai
Missed our webinar on Digital Asset Management and Personalization with AI? Don’t worry! Catch the recording here and learn how to drive more leads by using AI-based content organization and personalization. www.clarifai.com/events/webinar-digital-asset-management-…
#ai #contentorganization #digitalassetmanagement -
Clarifai
@clarifai
A huge thank you to Megan LaFlamme and Minh Tran for speaking at our webinar and to all those who joined us! We hope you found the session valuable for creating a hyper-personalized content experience.
#ai #marketing -
Clarifai
@clarifai
📅 Today’s the day! Join us at 1 pm ET to learn how AI can transform your lead pipeline with effective digital asset organization and personalization strategies. www.clarifai.com/events/webinar-ai-operating-model-for-co…
#ai #contentmarketing #marketing -
Clarifai
@clarifai
🔧 Drive more leads by delivering impactful content. Learn how to build and deploy AI for digital asset organization and personalization. Join us tomorrow. www.clarifai.com/events/webinar-ai-operating-model-for-co…
#marketing #ai #webinar -
Clarifai
@clarifai
⏳ Only 2 days left! Register for our webinar now to learn how to super-power your marketing with a solid AI operating model for content organization and personalization. www.clarifai.com/events/webinar-ai-operating-model-for-co…
#ai #webinar #aioperatingmodel #marketing -
Clarifai
@clarifai
📊 Learn how to easily tag, enrich, organize, and search your marketing content with AI. Drive more leads with a personalized digital experience. Register for our webinar on August 21st! www.clarifai.com/events/webinar-ai-operating-model-for-co…
#ai #contentmanagement -
Clarifai
@clarifai
👥 Meet our experts! Megan LaFlamme and Minh Tran will discuss how to bring together #marketing and #development teams to build an AI operating model for effective content organization and personalization. www.clarifai.com/events/webinar-ai-operating-model-for-co…
#ai #contentpersonalization #contentorganization -
Clarifai
@clarifai
Personalized content delivery can drive greater conversions. Join our webinar to learn how to build an AI operating model for content organization and personalization that can help you drive more leads! www.clarifai.com/events/webinar-ai-operating-model-for-co…
#ai #contentstrategy #marketing -
Clarifai
@clarifai
New release - Clarifai 10.7 is out now! ⚡️
We’re excited to introduce a new training template to fine-tune the latest Llama 3.1 model.
You can now use the template to fine-tune the Llama 3.1 8B instruct model within the Platform UI for extended context, instruction-following, or tasks such as text generation and text classification.
In addition to Llama 3.1 fine-tuning, this release also includes new features, new models (GPT-4o mini, Mistral Nemo, etc.), and other improvements, along with bug fixes.
Read more in the release blog here:
www.clarifai.com/blog/clarifai-10.7-finetune-llama-3.1 -
Clarifai
@clarifai
Drive more leads with a personalized content experience.
Register for our webinar to discover how to build an #ai operating model for #contentorganization and #personalization that can help you improve demand conversion. www.clarifai.com/events/webinar-ai-operating-model-for-co…
#marketing #contentpersonalization -
Clarifai
@clarifai
Mistral NeMo 12B is the latest and best small model from @MistralAI, built in collaboration with NVIDIA. This state-of-the-art 12B model with a 128k context length. 🔥
The model offers a balance between size and performance, providing advanced capabilities on a manageable scale.
It excels in reasoning, world knowledge, and coding tasks, making it one of the most capable models within its size category.
The Instruct variant is now available on the Clarifai Platform.
Try out the model here or access it with an API:
clarifai.com/mistralai/completion/models/mistral-nemo-ins… -
Clarifai
@clarifai
📅 Save the date!
On August 21st, Megan LaFlamme and Minh Tran will share how to overcome challenges in building AI projects for content organization and personalization. Don’t miss it! www.clarifai.com/events/webinar-ai-operating-model-for-co…
#contentmarketing #ai #contentorganization -
Clarifai
@clarifai
🚀 Webinar alert!
Join us on Aug 21 to learn how to build an AI operating model for content organization and personalization. Learn more about the webinar here - www.clarifai.com/events/webinar-ai-operating-model-for-co…
#ai #webinar #personalization #contentorganization -
Clarifai
@clarifai
Last week, we hosted a webinar on operationalizing and orchestrating AI.
If you missed it, don’t worry! Get the recording here and learn more about the best practices to move your AI projects from prototype to production faster.
www.clarifai.com/events/webinar-ai-prototype-to-production -
Clarifai
@clarifai
"People can quickly get running on a laptop. There’s a bunch of tools that either are open source or provided by companies that help you stitch [it all] together, but at the end of the day, those are really just API clients. " Great interview with @mattzeiler in @Forbes about how companies get past the "false finish line" of #ai prototypes with AI workflow orchestration.
www.forbes.com/sites/meganpoinski/2024/07/25/how-cios-can… -
Clarifai
@clarifai
📺 Missed our webinar on operationalizing and orchestrating AI? Don’t worry! Catch the recording here and learn more about the best practices to move your AI projects from prototype to production faster. www.clarifai.com/events/webinar-ai-prototype-to-productio…
-
Clarifai reposted
Sumanth
@Sumanth_077
Build a LLM app with RAG to chat with PDF files using Llama 3.1 8B instruct in just 4 lines of Python code:
-
Clarifai reposted
Clarifai
@clarifai
Access Llama 3.1 with an API!
Meta AI released Llama 3.1 405B, the first openly available model that rivals the top AI models in state-of-the-art capabilities in general knowledge, steerability, math, tool use, and multilingual translation.
Along with the 405B variant, Meta also released the upgraded versions of the 8B and 70B models.
These models are multilingual and have a significantly longer context length of 128K, state-of-the-art tool use, and overall stronger reasoning capabilities.
The 8B instruct variant is now available on the Clarifai Platform.
Chat with the model and access it via API here: clarifai.com/meta/Llama-3/models/llama-3_1-8b-instruct -
Clarifai
@clarifai
🙏 A huge thank you to our incredible hosts, Pete Pacent and Megan LaFlamme, and to everyone who joined our webinar on "AI prototype to production: Operationalizing and Orchestrating AI". We hope you found it insightful and valuable.
-
Clarifai
@clarifai
⏰ Just 1 hour to go! Ready to move past AI prototypes? Join our experts Pete Pacent and Megan LaFlamme.
www.clarifai.com/events/webinar-ai-prototype-to-productio… -
Clarifai
@clarifai
Access Llama 3.1 with an API!
Meta AI released Llama 3.1 405B, the first openly available model that rivals the top AI models in state-of-the-art capabilities in general knowledge, steerability, math, tool use, and multilingual translation.
Along with the 405B variant, Meta also released the upgraded versions of the 8B and 70B models.
These models are multilingual and have a significantly longer context length of 128K, state-of-the-art tool use, and overall stronger reasoning capabilities.
The 8B instruct variant is now available on the Clarifai Platform.
Chat with the model and access it via API here: clarifai.com/meta/Llama-3/models/llama-3_1-8b-instruct -
Clarifai
@clarifai
⏳ Only 1 day left! Don’t miss out on expert insights from Pete Pacent and Megan LaFlamme on operationalizing AI. www.clarifai.com/events/webinar-ai-prototype-to-productio…
#ai #webinar -
Clarifai
@clarifai
Introducing GPT-4o mini, the smallest, intelligent, and the cost-efficient model from OpenAI, which is now available on the Clarifai Platform. 🎉
- The model surpasses GPT-3.5 Turbo and other small models on academic benchmarks across both textual intelligence and multimodal reasoning and is 60% cheaper than GPT-3.5 Turbo.
- The model comes with a context window of 128K tokens and supports up to 16K output tokens per request.
- Supports text and vision via the API.
GPT-4o mini is now available on the Clarifai Platform. You can access it using an API with just a few lines of code and integrate it into your applications:
Try it out:
clarifai.com/openai/chat-completion/models/gpt-4o-mini -
Clarifai
@clarifai
🧠 Move past the “false finish line” of AI prototypes. Join our webinar on operationalizing and orchestrating AI to discover strategies for successful AI implementation. www.clarifai.com/events/webinar-ai-prototype-to-productio…
#ai #digitaltransformation -
Clarifai
@clarifai
📊 Struggling to move your AI projects from prototype to production? You're not alone. Join our webinar on July 24 at 1 pm ET to learn best practices for operationalizing and orchestrating AI. Sign up now!
#ai #tech www.clarifai.com/events/webinar-ai-prototype-to-productio… -
Clarifai
@clarifai
Introducing the new Auto Annotation feature: Automatically annotate all image inputs in your dataset with just one click, regardless of their size.
Check out the complete tutorial here:
youtu.be/CD25FaeqVRc -
Clarifai
@clarifai
Exciting news! @clarifai has been named a Gartner #samplevendor in the 2024 Hype Cycle™ for Enterprise Architecture report. Check out the report here - www.clarifai.com/report/clarifai-gartner-hype-cycle?hss_c… 🎉
#ai #enterprisearchitecture -
Clarifai
@clarifai
🔍 Did you know 61% of CIOs say their security infrastructure isn’t AI-ready? Learn how to overcome this and other challenges to Productionize your AI projects in our upcoming webinar.
www.clarifai.com/events/webinar-ai-prototype-to-productio…
#ai #security #compliance -
Clarifai reposted
Clarifai
@clarifai
Multistage RAG Pipeline With DSPy!
Leverage Clarifai and DSPy to build a multi-stage RAG module.
The module extracts keywords to broaden the search, retrieves information using both the query and keywords, and employs a separate language model to evaluate and rank the retrieved passages based on their relevance to the user's question.
This multi-stage approach with distinct components allows for a more comprehensive and potentially more accurate information retrieval process.
Check out the detailed tutorial here:
docs.clarifai.com/integrations/DSPy/multistage-dspy/ -
Clarifai
@clarifai
Multistage RAG Pipeline With DSPy!
Leverage Clarifai and DSPy to build a multi-stage RAG module.
The module extracts keywords to broaden the search, retrieves information using both the query and keywords, and employs a separate language model to evaluate and rank the retrieved passages based on their relevance to the user's question.
This multi-stage approach with distinct components allows for a more comprehensive and potentially more accurate information retrieval process.
Check out the detailed tutorial here:
docs.clarifai.com/integrations/DSPy/multistage-dspy/ -
Clarifai reposted
Sumanth
@Sumanth_077
Build a LLM app with RAG to chat with PDF files using Claude 3.5 Sonnet in just 4 lines of Python code:
-
Clarifai
@clarifai
👥 Join Clarifai's Pete Pacent and Megan LaFlamme for a webinar on July 24 as they share best practices for deploying AI at scale. Don’t miss out on this opportunity!
www.clarifai.com/events/webinar-ai-prototype-to-productio…
#ai #innovation -
Clarifai
@clarifai
💡 New Tutorial 💡
Learn how to auto-annotate your image datasets with a single click, regardless of their size, using the Clarifai Platform.
The platform now supports the Auto Annotation feature, allowing you to leverage machine learning models and workflows from the community to automatically annotate your data.
Check out the full tutorial here:
youtu.be/CD25FaeqVRc -
Clarifai
@clarifai
🎉 Clarifai is now integrated with EmbedChainAI! 🚀
EmbedChain is an open-source framework designed to streamline the integration of LLMs with custom knowledge bases.
With this integration, you can now leverage the LLMs and embedding models from the Clarifai Platform to build RAG applications.
Check out this RAG tutorial to get started: docs.clarifai.com/integrations/embedchain/embedchain-tuto… -
Clarifai
@clarifai
New release - Clarifai 10.6 is out now! ⚡️
We are excited to introduce the new Auto Annotation feature, that lets you auto-annotate entire image datasets with a single click, regardless of their size, using the Clarifai Platform.
You can leverage machine learning models and workflows from the Community to automatically annotate the data.
Additionally, you can determine which concepts to annotate automatically and which to review manually based on defined confidence thresholds, reviewing only those inputs that fall below your specified accuracy thresholds.
In addition to Auto Annotation, the release also includes new features, the latest Clarifai Embedchain integration, new models (Claude 3.5 Sonnet, Florence-2), and other improvements, alongside bug fixes.
Read more in the release blog here:
www.clarifai.com/blog/clarifai-10.6-click-annotate-domina… -
Clarifai
@clarifai
🚀 Are your AI projects stuck in prototype? Join us on July 24th to learn how to operationalize and orchestrate AI from prototype to production. Register now!
www.clarifai.com/events/webinar-ai-prototype-to-productio…
#ai #webinar -
Clarifai
@clarifai
🤔 Is your organization struggling to build AI at scale?
The journey from prototypes to production is filled with challenges. Our upcoming webinar on “AI Prototype to Production” will help you build AI faster and at scale. Join us.
www.clarifai.com/events/webinar-ai-prototype-to-productio… -
Clarifai
@clarifai
Build RAG applications using the latest Claude 3.5 Sonnet model from @AnthropicAI in just 4 lines of code!
Building a RAG system is simple with the Clarifai Python SDK. Here are the steps:
• Install the Clarifai Python SDK, Llama-index, and set your Clarifai Personal Access Token as an environment variable.
• Import the RAG class and set up the RAG system by passing your Clarifai user ID and the Claude 3.5 Sonnet model URL.
• Upload the data.
• Chat with your data.
Access Claude 3.5 Sonnet model from the community here:
clarifai.com/anthropic/completion/models/claude-3_5-sonnet -
Clarifai
@clarifai
Create datasets and upload the data using the Python SDK!
You can easily create datasets and upload the data to your Clarifai app using the Python SDK.
Below is an example of uploading image data to your existing Clarifai App.
Check this out: docs.clarifai.com/sdk/managing-datasets -
Clarifai
@clarifai
Batch Predict using the Python SDK! 🚀
Process multiple inputs in a single request by leveraging Clarifai's Predict API batch prediction feature.
This allows you to streamline the prediction process, saving time and resources.
The example below demonstrates how to submit image inputs in a batch to the General Image Recognition model.
Check this out: 👇
docs.clarifai.com/sdk/Inference-from-AI-Models/Advance-In… -
Clarifai
@clarifai
Do LLMs Reign Supreme in Few-Shot NER?
The following blog post discusses the details of using LLMs, especially open-source ones like Llama-2, for few-shot NER tasks and their challenges.
Read the blog post here: www.clarifai.com/blog/do-llms-reign-supreme-in-few-shot-n…
Checkout the code here: -
Clarifai
@clarifai
Moving past the false finish line of AI prototypes is hard, but Clarifai makes it easier!
Join Pete Pacent (Head of Product) and Megan LaFlamme (Head of Product Marketing) for a live webinar on July 24 to learn how to move past the false finish line with AI prototypes and into production and the best practices for operationalizing and orchestrating AI.
Register here:
www.clarifai.com/events/webinar-ai-prototype-to-production -
Clarifai
@clarifai
Object Detection Using Florence-2 🔥
The recently released Florence-2 model demonstrates strong zero-shot capabilities across tasks such as captioning, object detection, grounding, and segmentation.
Below is an example of using the model for an object detection task and getting the bounding box coordinates for an image.
The model is available on the Clarifai Platform. Try it out: clarifai.com/microsoft/florence/models/florence-2-large
#objectdetection #computervision #vllm -
Clarifai
@clarifai
Claude 3.5 Sonnet has been released and is now available on the Clarifai Platform!
Claude 3.5 Sonnet is the first release in the @AnthropicAI 3.5 model family and outperforms competitor models on key evaluations.
Key takeaways of the model:
- Claude 3.5 Sonnet is 2x faster than Claude 3 Opus.
- Claude 3.5 Sonnet is now the strongest vision model in the Claude family, surpassing Claude 3 Opus across all standard vision benchmarks.
- Claude 3.5 Sonnet sets new industry benchmarks for graduate-level reasoning (GPQA), undergraduate-level knowledge (MMLU), and coding proficiency (HumanEval).
Try out the model now on the Clarifai Platform:
clarifai.com/anthropic/completion/models/claude-3_5-sonnet -
Clarifai
@clarifai
Florence-2 is out and is now available on the Clarifai Platform! 🔥
Florence-2 is the new lightweight vision-language model open-sourced by Microsoft.
Here are the key takeaways of the model:
• Handles a variety of vision and vision-language tasks through a prompt-based approach.
• Trained on the extensive FLD-5B dataset.
• Available in two variants: Florence-2-base (0.23B parameters) and Florence-2-large (0.77B parameters).
Try out the Florence-2 Large now on Clarifai: 👇
clarifai.com/microsoft/florence/models/florence-2-large -
Clarifai
@clarifai
We are thrilled to announce Clarifai has been assessed as "Awardable" for Department of Defense work in the Chief Digital and Artificial Intelligence Office’s (CDAO) Tradewinds Solutions Marketplace.
This recognition highlights our leading role in developing AI solutions for Government and Defense applications.
Read the announcement here:
www.clarifai.com/press-release/clarifai-assessed-awardabl… -
Clarifai
@clarifai
Ingest Email Messages from Salesforce to the Clarifai App using the Clarifai @UnstructuredIO integration! 🔥
The following tutorial demonstrates how to ingest email messages from Salesforce to the Clarifai app and then build a RAG system to classify, and summarize emails.
Check this out:
docs.clarifai.com/integrations/unstructured/salesforce -
Clarifai
@clarifai
Transform your Content Creation with the Content Generation Template!🚀
The Content Generation Template is an app template that includes several ready-to-use workflows for content creation, leveraging different LLM models optimized through various prompt engineering techniques.
The workflows are optimized for various use cases such as email writing, blog writing, social media content, and more.
Access and clone the template to get started:
clarifai.com/clarifai/content-generation-template -
Clarifai
@clarifai
Leverage AI to Assist and Optimize the Process of Annotating Data! 🚀
You can request predictions from any model or workflow available to you for a particular input, and then review, correct, or validate them before converting them into annotations.
AI-Assist provides several key benefits, including:
• Significantly accelerating the labelling process, reducing the time required to create labelled datasets.
• Reducing labor costs associated with manual labelling through automation.
• Providing consistent labelling with AI models, minimizing human errors.
• Allowing efficient and scalable handling of large datasets or rapidly changing data.
Check out the guide on AI-assisted labelling in this video: youtu.be/x4lC5sz-Oqs -
Clarifai
@clarifai
New release - Clarifai 10.5 is out now! ⚡️
We are excited to introduce Fine-Tuning Large Language Models using the Clarifai Platform.
Fine-tuning allows you to adapt foundational text-to-text models to specific tasks or domains, making them more suitable for particular applications.
You can now select a pre-configured model template from the platform to train on your data, such as Llama2 7/13B or Mistral models with GPTQ-Lora.
In addition to LLM fine-tuning, the release also includes the new Coding App Template, LiteLLM Integration, New models (GPT-4o, Gemini 1.5 Flash, Snowflake Arctic-Instruct), and other improvements, along with bug fixes.
Read more in the release blog here: www.clarifai.com/blog/clarifai-10.5-gear-up-your-ai-fine-…
#llmfinetuning -
Clarifai
@clarifai
Cross-modal search with Clarifai!
Cross-modal search lets you query diverse data types, such as text, images, audio, and video.
It breaks down the barriers between different data modalities, offering a more holistic and intuitive search experience.
The notebook below provides an introductory guide to setting up a cross-modal search system on the Clarifai platform using the Python SDK.
It uses a text-to-image search example to illustrate how quickly such a system can be set up using only a few lines of code.
Notebook:👇
github.com/Clarifai/examples/blob/main/search/cross_modal… -
Clarifai
@clarifai
RAG using Google's Gemini 1.5 Flash in just 4 lines of code!
Gemini 1.5 Flash is a cost-efficient, high-speed foundation LLM optimized for multimodal tasks, ideal for applications requiring rapid processing and scalability
Building a RAG system is simple with the Clarifai Python SDK. Here are the steps:
• Install the Clarifai Python SDK, Llama-index, and set your Clarifai Personal Access Token as an environment variable.
• Import the RAG class and set up the RAG system by passing your Clarifai user ID and the Gemini-1.5 Flash model URL.
• Upload the data.
• Chat with your data.
Access Gemini 1.5 Flash model from the community here: clarifai.com/gcp/generate/models/gemini-1_5-flash
#rag #geminiflash -
Clarifai
@clarifai
The Clarifai Node.js SDK offers a comprehensive set of tools to integrate Clarifai's AI platform capabilities into your applications, leveraging its computer vision and natural language processing capabilities.
With just a few lines of code, you can utilize cutting-edge artificial intelligence to unlock valuable insights from visual and textual content.
Check out the installation guide here: docs.clarifai.com/sdk/node-js-installation/
Github repo: github.com/Clarifai/clarifai-nodejs -
Clarifai
@clarifai
Great discussions yesterday during the first day of the @EdgeComputingEx! Clarifai's Head of Product, Pete Pacent, participated in an excellent panel about how visionary companies are using AI at the edge to transform industries like Energy and Automatives.
Excellent conversation with Uber, Fisker, National Grid, and ZEDEDA about how development teams and companies can address concerns around edge security and AI, model performance and distillation, and moving from prototype to production.
Are you using AI at the edge? What are some of the biggest challenges your organization is facing around deploying AI at scale?
A big thanks to everyone who attended and contributed to the conversations and @TechEx_Event for organizing this event! Enjoy day 2!
#techex #edgecomputingexpo -
Clarifai
@clarifai
In the race to harness the value of AI, most organizations are getting stuck moving AI from prototype to production. AI prototypes can be highly effective, and even cost efficient, but can create a “false finish line” that production-ready deployments are around the corner.
Based on our learnings as a pioneer of the full-stack AI platform and accelerating AI development for over a decade with the largest organizations and governments, we have compiled this new guide on “Orchestrating and Operationalizing AI” that addresses the top challenges in deploying AI at scale and outlines key pillars needed to adopt AI technologies, optimize value, and mitigate risk.
Read the announcement to learn more and to access the complete guide:
www.clarifai.com/press-release/clarifai-unveils-comprehen… -
Clarifai
@clarifai
Here is a summary of all the latest LLMs and multi-modal LLMs hosted and integrated into the Clarifai Community in May, along with links to try them out and access via an API! 💥
• Codestral-22B-Instruct: Codestral is an open-weight, 22B parameter model from Mistral AI with a 32K context window explicitly designed for code generation tasks.
clarifai.com/mistralai/completion/models/codestral-22b-in…
• MiniCPM-Llama3-V 2.5: A high-performance, efficient 8B parameter multimodal model excelling in OCR, multilingual support, and multimodal tasks.
clarifai.com/openbmb/miniCPM/models/miniCPM-Llama3-V-2_5
• GPT-4o (omni): GPT-4o is a multimodal AI model from Open AI that excels in processing and generating text, audio, and images.
The model offers rapid response times and improved performance across languages and tasks.
clarifai.com/openai/chat-completion/models/gpt-4o
• Gemini-1.5-Flash: Gemini 1.5 Flash is a cost-efficient, high-speed foundational LLM optimized for multimodal tasks, ideal for applications requiring rapid processing and scalability.
clarifai.com/gcp/generate/models/gemini-1_5-flash
• Snowflake-Arctic Instruct: Cost-effective, enterprise-focused open LLM that excels in SQL, coding, and instruction-following from Snowflake. The model is designed to be both cost-effective and powerful, providing state-of-the-art performance for enterprise applications.
clarifai.com/snowflake/arctic/models/snowflake-arctic-ins…
• Openchat-3.6-8b: A high-performance, open-source LLM fine-tuned from llama3-8b with C-RLFT, delivering ChatGPT-level performance across various tasks.
clarifai.com/openchat/openchat/models/openchat-3_6-8b-202…
• CogVLM-Chat: An advanced open-source visual language model that can process and understand both visual and textual data. Compared to BLIP-2, Otter, and various LLaMA variants, CogVLM-Chat not only surpasses in scoring but also shows comprehensive strength in handling adversarial and complex scenarios.
clarifai.com/thudm/cogvlm/models/cogvlm-chat
• Qwen-VL-Chat: A high-performing Large Vision Language Model (LVLM) by Alibaba Cloud for text-image dialogue tasks, excelling in zero-shot captioning, VQA, and referring expression comprehension while supporting multilingual dialogue.
clarifai.com/qwen/qwen-VL/models/qwen-VL-Chat
• Mixtral-8x22B-instruct: The latest and largest mixture of expert LLM from Mistral AI using a mixture of 8 experts (MoE) 22b models. The model comes with multilingual support and can handle a context of 32k tokens.
clarifai.com/mistralai/completion/models/mixtral-8x22B-In…
Qwen1.5-110B-Chat: 100 billion parameter model, demonstrates competitive performance in base language tasks and significant improvements in chatbot evaluations and boasts multilingual capabilities.
clarifai.com/qwen/qwenLM/models/qwen1_5-110B-chat
Try them out! -
Clarifai
@clarifai
AI is changing how the world runs, and organizations face the challenge of navigating this transformative journey quickly and efficiently.
Clarifai can help. Introducing a new Operating Model for Orchestrating & Operationalizing Enterprise AI!
If this is your organization, let's talk!
www.clarifai.com/report/an-operating-model-for-orchestrat… -
Clarifai
@clarifai
.@MistralAI introduced Codestral, their first-ever code model and the model is now available on the Clarifai Platform! 🔥
Codestral is an open-weight, 22B parameter generative AI model that comes with a 32K context window explicitly designed for code generation tasks. It helps developers write and interact with code through a shared instruction and completion API endpoint.
Codestral is trained on a diverse dataset of 80+ programming languages, including the most popular ones such as Python, Java, C, C++, JavaScript, and Bash.
The instruct variant of the Codestral model is now available on the Clarifai Platform.
Try it out or access it with an API here.👇
clarifai.com/mistralai/completion/models/codestral-22b-in… -
Clarifai
@clarifai
Build RAG with Clarifai using @getpostman! 🚀
You can access the Clarifai platform to create RAG applications with ease using just the APIs.
The steps include:
1. Create a Clarifai app with “Text” as the base workflow.
2. Create a RAG prompter operator model and then create the model version.
3. Create a RAG Prompter Workflow along with a text-to-text node and specify your preferred LLM.
4. Upload your data to the app created above using the Python SDK.
Check out the guide on how to upload the data: github.com/Clarifai/clarifai-python?tab=readme-ov-file.
5. Make API predictions to the created RAG prompter workflow using Postman.
Check out the public Postman collection here that will walk you through these steps: 👇
documenter.getpostman.com/view/30622694/2s9YkuZdro -
Clarifai
@clarifai
MiniCPM-Llama3-V 2.5 is now available on the Clarifai Platform! 🎉
MiniCPM-Llama3-V 2.5 is a high-performance, efficient 8B parameter multimodal model excelling in OCR, multilingual support, and multimodal tasks.
Here are some key capabilities of the model:
• Leading Performance: Achieves an average score of 65.1 on OpenCompass, outperforming larger proprietary models like GPT-4V-1106, Gemini Pro, Claude 3, and Qwen-VL-Max.
• Strong OCR Capabilities: Excels in OCR tasks, handling images with high pixel counts and various aspect ratios, scoring over 700 on OCRBench.
• Trustworthy Behavior: Utilizes the latest RLAIF-V method to minimize hallucination rates, achieving a 10.3% rate on Object HalBench.
• Multilingual Support: Extends capabilities to over 30 languages, leveraging Llama 3's multilingual strengths and VisCPM's cross-lingual generalization.
• Efficient Deployment: Implements model quantization and optimizations for CPUs and NPUs, ensuring efficient performance on edge devices, including mobile phones with Qualcomm chips.
The model is now available on the platform. Try out the model and access it with an API here: 👇
clarifai.com/openbmb/miniCPM/models/miniCPM-Llama3-V-2_5 -
Clarifai reposted
Sumanth
@Sumanth_077
Build a LLM app with RAG to chat with PDF files using Llama-3 in just 4 lines of Python code (step-by-step instructions):
-
Clarifai
@clarifai
We are excited to introduce the new Clarifai - LiteLLM integration! 🎉🥳
@LiteLLM is a library for seamless integration with various LLM provider's APIs that standardizes interactions through an OpenAI API format.
This integration allows you to call various open-source and third-party LLMs available on the Clarifai Platform using the same input/output format.
Check out the guide here to get started:
github.com/BerriAI/litellm/blob/main/docs/my-website/docs… -
Clarifai
@clarifai
Introducing the Coding App template! 🎉🔥
App templates are pre-built, ready-to-use apps that streamline the app creation process, making it faster and more efficient. 🚀
The Coding App template explores various coding scenarios and includes pre-built workflows tailored to address different coding use cases, employing diverse models specialized for each unique situation.
There are various workflows specialized for Code Completion, Code Generation, Code Summarization, Code Refactoring and Code Debugging.
Check this out and get started by cloning the app template:
clarifai.com/clarifai/coding-template
#apptemplates -
Clarifai
@clarifai
RAG using GPT-4o (omni) in just 4 lines of code!
GPT-4o is a multimodal AI model from Open AI that offers rapid response times and improved performance across languages and tasks while incorporating advanced safety features.
Building a RAG system is simple with the Clarifai Python SDK. Here are the steps:
• Install the Clarifai Python SDK, Llama-index, and set your Clarifai Personal Access Token as an environment variable.
• Import the RAG class and set up the RAG system by passing your Clarifai user ID and passing the GPT-4o model URL.
• Upload the data.
• Chat with your data.
Access GPT-4o here: clarifai.com/openai/chat-completion/models/gpt-4o -
Clarifai
@clarifai
A Gentle Introduction to Retrieval Augmented Generation (RAG) for the Intelligence Community!
RAG has revolutionized how we access and utilize data, offering real-time information retrieval to enhance language models.
Dive deep into this blog on the importance of RAG, especially for intelligence analysts, with a real-world use case.
Read here: intelligencecommunitynews.com/ic-insiders-a-gentle-introd… -
Clarifai
@clarifai
Google announced Gemini 1.5 Flash yesterday, and the model is now available on the Clarifai Platform! 🎉
Gemini 1.5 Flash is a cost-efficient, high-speed foundational LLM optimized for multimodal tasks, ideal for applications requiring rapid processing and scalability.
The model is lighter-weight than 1.5 Pro and designed to be fast and efficient for serving at scale.
Additionally, the model excels at summarization, chat applications, image and video captioning, data extraction from long documents and tables, and more.
The model is now available on the Clarifai platform.
Try out the model here: 👇
clarifai.com/gcp/generate/models/gemini-1_5-flash
#geminiflash -
Clarifai
@clarifai
🎉 GPT-4o (omni) is now available on the Clarifai Platform! 🎉
Yesterday, OpenAI released GPT-4o, a multimodal AI model that excels in processing and generating text, audio, and images.
The model offers rapid response times and improved performance across languages and tasks while incorporating advanced safety features.
It can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in a conversation.
The model also matches GPT-4 Turbo performance on text in English and code, with significant improvement in non-English languages, while also being much faster and 50% cheaper in the API.
GPT-4o is also better at vision and audio understanding compared to existing models.
Text and image inputs are now available via API.
You can try out the model now on the Clarifai Platform and access it via API here: 👇
clarifai.com/openai/chat-completion/models/gpt-4o
#gpt4o -
Clarifai
@clarifai
New release - Clarifai 10.4 is out now! ⚡️
We are excited to introduce the new App templates.
App templates are pre-built, ready-to-use apps that streamline the app creation process, making it faster and more efficient. 🚀
The Clarifai platform provides app templates for applications such as RAG, document summarization, image moderation, content generation, sentiment analysis, and others to help you hit the ground running.
In addition to the newly introduced templates, the release also includes new features, new models such as Llama-3, Mixtral, Command R Plus, performance and UX improvements, and bug fixes.
Read more in the release blog here: clarifai.com/blog/clarifai-10.4
#apptemplates -
Clarifai
@clarifai
Streamline and optimize the data processing pipelines for LLMs with the Clarifai and @UnstructuredIO integration! 🚀
Unstructured IO provides libraries with open-source components for ingesting and pre-processing data.
The following notebook is a step-by-step guide on how to utilize Clarifai as your destination connector using unstructured and import raw files from the S3 bucket.
Check out the Notebook here:
github.com/Clarifai/examples/blob/main/Integrations/Unstr… -
Clarifai
@clarifai
Introducing Qwen-VL-Chat, a high-performing Large Vision Language Model (LVLM) by Alibaba Cloud.
The model is most suitable for text-image dialogue tasks, excelling in zero-shot captioning, VQA, and referring expression comprehension while supporting multilingual dialogue.
Try out the model now on the Clarifai Platform here:
clarifai.com/qwen/qwen-VL/models/qwen-VL-Chat
#lvlm -
Clarifai
@clarifai
Introducing CogVLM-Chat, an advanced open-source visual language model that can process and understand both visual and textual data.
The model utilizes the foundational architecture of CogVLM-17B.
When compared to other models like BLIP-2, Otter, and various LLaMA variants, CogVLM-Chat not only surpasses in scoring but also shows comprehensive strength in handling adversarial and complex scenarios.
The model is now available on the Clarifai Platform!
Try it out or access with an API here: clarifai.com/thudm/cogvlm/models/cogvlm-chat -
Clarifai reposted
Clarifai
@clarifai
Build Retrieval Augmented Generation(RAG) using Llama-3 in just 4 lines of code:
-
Clarifai
@clarifai
Build Retrieval Augmented Generation(RAG) using Llama-3 in just 4 lines of code:
Post not available here, view on X
Clarifai
@clarifai
Here is the step-by-step tutorial on building RAG in 4 lines of code.
Checkout the video: youtu.be/G6OsFovu-9Q -
Clarifai reposted
Clarifai
@clarifai
Introducing Snowflake-Arctic Instruct, a new cost-effective, enterprise-focused LLM that excels in SQL, coding, and instruction-following from @SnowflakeDB! 🔥
The model is designed to be both cost-effective and powerful, providing state-of-the-art performance for enterprise applications.
Available now on the Clarifai Platform:
Try it out: ⬇️
clarifai.com/snowflake/arctic/models/snowflake-arctic-ins… -
Clarifai
@clarifai
Introducing Snowflake-Arctic Instruct, a new cost-effective, enterprise-focused LLM that excels in SQL, coding, and instruction-following from @SnowflakeDB! 🔥
The model is designed to be both cost-effective and powerful, providing state-of-the-art performance for enterprise applications.
Available now on the Clarifai Platform:
Try it out: ⬇️
clarifai.com/snowflake/arctic/models/snowflake-arctic-ins… -
Clarifai
@clarifai
Clarifai at GEOINT 2024! 🚀
We're eagerly looking forward to connecting at #geoint2024 this year.
See Clarifai's AI platform in action and applied to real-world government case studies.
We are bringing folks from across the company to answer questions, build partnerships, deliver demos, and presentations.
Featured use cases include:
• AI-assisted and automated data labelling and annotation
• Document and media intelligence with RAG
• People tracking with FMV
• AI-Assisted Document Classification
• Leveraging new LLMs
📅 May 6-8
📍Stop by to meet the team at booth #1902 to see some cool use cases:
Read more here: www.clarifai.com/events/geoint2024 -
Clarifai reposted
Clarifai
@clarifai
Building RAG applications with Clarifai using @getpostman!
If you are around tomorrow, come join @mattzeiler, CEO and Founder of Clarifai, as he will be taking the stage as a speaker at the Postman annual conference, POST/CON 24. 🎉
Don't miss the opportunity to learn how to build RAG applications using Postman APIs, and how AI development on a full-stack AI platform simplifies how teams create, share, and run AI at scale.
For more details, visit the POST/CON site here:
www.postman.com/postcon/
#postcon -
Clarifai
@clarifai
Building RAG applications with Clarifai using @getpostman!
If you are around tomorrow, come join @mattzeiler, CEO and Founder of Clarifai, as he will be taking the stage as a speaker at the Postman annual conference, POST/CON 24. 🎉
Don't miss the opportunity to learn how to build RAG applications using Postman APIs, and how AI development on a full-stack AI platform simplifies how teams create, share, and run AI at scale.
For more details, visit the POST/CON site here:
www.postman.com/postcon/
#postcon -
Clarifai
@clarifai
We're excited to announce that Clarifai will be at GEOINT 2024! 🥳
See Clarifai's AI platform in action and applied to real-world government case studies through demos, presentations, and discussions with Clarifai specialists.
Featured use cases include:
• AI-assisted and automated data labelling and annotation
• Document and media intelligence with RAG
• People tracking with FMV
• AI-Assisted Document Classification
• Leveraging new LLMs
📅 May 6-8
📍 Visit Booth #1902
www.clarifai.com/events/geoint2024
#geoint2024 -
Clarifai reposted
Clarifai
@clarifai
Introducing the Clarifai Node.js SDK [Developer Preview] 📦
The Node.js SDK offers a comprehensive set of tools to integrate Clarifai's AI platform and leverage computer vision and natural language capabilities into your applications.
With just a few lines of code, you can leverage cutting-edge AI to extract valuable insights from visual and textual content.
Below is a quick example of how to use the celebrity face recognition model to predict the celebrity in a given picture.
Check out the documentation here:
docs.clarifai.com/nodejs-sdk/installation-guide/
#clarifai #nodejs -
Clarifai
@clarifai
Introducing the Clarifai Node.js SDK [Developer Preview] 📦
The Node.js SDK offers a comprehensive set of tools to integrate Clarifai's AI platform and leverage computer vision and natural language capabilities into your applications.
With just a few lines of code, you can leverage cutting-edge AI to extract valuable insights from visual and textual content.
Below is a quick example of how to use the celebrity face recognition model to predict the celebrity in a given picture.
Check out the documentation here:
docs.clarifai.com/nodejs-sdk/installation-guide/
#clarifai #nodejs -
Clarifai
@clarifai
Interact seamlessly with Llama 3, Mixtral-8x22B, Gemini-1.5 Pro, Command R+, Claude 3, and many other open-source and third-party models with the new UI on the Clarifai Platform! ⚡️🚀
The following video guides you through the new UI, showcasing how you can configure the inference parameters, access these models via an API, and effortlessly generate or convert text based on a given input prompt.
Watch the video:
youtu.be/n2g0UihfOkE -
Clarifai reposted
Clarifai
@clarifai
🎉 Try out Llama-3 70B and Llama-3 70B Instruct models on the Clarifai Platform! 🚀
Last week, we announced the availability of the Llama-3 8B instruct model on the platform, and now we are excited to announce two new models from the Llama-3 family.
Llama-3 70B: clarifai.com/meta/Llama-3/models/llama-3-70B
Llama-3 70B Instruct: clarifai.com/meta/Llama-3/models/llama-3-70B-Instruct
Try them out and easily access them via an API!
#llama3 -
Clarifai
@clarifai
🎉 Try out Llama-3 70B and Llama-3 70B Instruct models on the Clarifai Platform! 🚀
Last week, we announced the availability of the Llama-3 8B instruct model on the platform, and now we are excited to announce two new models from the Llama-3 family.
Llama-3 70B: clarifai.com/meta/Llama-3/models/llama-3-70B
Llama-3 70B Instruct: clarifai.com/meta/Llama-3/models/llama-3-70B-Instruct
Try them out and easily access them via an API!
#llama3 -
Clarifai
@clarifai
BIG NEWS: @AIatMeta has released Llama 3, the most capable openly available LLM, and we are excited to announce that the model is now available on the Clarifai Platform! 🎉 🔥
Here are all the details you need to know: 🚀
• Trained on more than 15T tokens, 7x+ larger than Llama 2's dataset.
• 8B and 70B parameters are openly available pre-trained and fine-tuned models.
• Improved tokenizer with a vocabulary of 128K tokens for better performance.
• New capabilities, including enhanced reasoning and coding.
• 3x more efficient training than Llama 2.
Try out the llama3-8B Instruct model on the platform, or access it with an API here: 👇
clarifai.com/meta/Llama-3/models/Llama-3-8B-Instruct
#llama3 #llms -
Clarifai
@clarifai
.@MistralAI just announced Mistral 8x22B, their latest open model and we are excited to announce that the model is now available on the Clarifai Platform! 🎉
Key details of Mixtral 8x22B:
• 64K token context window
• Outperforms all open models
• Fluent in English, French, Italian, German, and Spanish
• Strong math and coding capabilities
• Capable of function calling
Try out the model here:
clarifai.com/mistralai/completion/models/mixtral-8x22B
#mixtral8x22b -
Clarifai
@clarifai
We're thrilled to announce that @mattzeiler, CEO and Founder of Clarifai, will be taking the stage as a speaker at the @getpostman annual conference POST/CON 24 in San Francisco on May 1, 2024. 🎉
Join us to learn how to Build AI Faster with APIs and Public Workspaces, as Matthew Zeiler walks through how AI development on a full-stack AI platform simplifies how teams create, share, and run AI at scale.
For more details and registration, visit the POST/CON site here: www.postman.com/postcon/
#postcon -
Clarifai
@clarifai
If you're at New York University tomorrow, don't miss the chance to hear from Douglas Shapiro, Chief of Staff & Senior Director of Operations at Clarifai, as he participates in the Research Symposium on "Bridging AI Innovation and Societal Impact."
Doug will be featured on a panel discussion focusing on AI Ecosystems: Opportunities for Cross-Sector Interactions and Collaborations.
#ai #nyu #researchsymposium -
Clarifai
@clarifai
Clarifai CEO @mattzeiler will be at this year’s @Jefferies Private Internet Conference on April 16-17th, engaging in discussions surrounding 'In the Age of AI' with technology thought leaders and investors.
#jefferiesprivateinternet -
Clarifai
@clarifai
Google's most capable generative AI model, Gemini 1.5 Pro, is now available on the Clarifai Platform! 🎉
One of the key differentiators of this model is its incredibly long context capabilities, supporting the standard 128,000 context window and capable of processing up to 1 million tokens.
The model is multimodal and can process and understand information from various modalities, including text, code, images, audio, and video.
Try it here: clarifai.com/gcp/generate/models/gemini-1_5-pro -
Clarifai reposted
Unstructured
@UnstructuredIO
New destination connector! 🔥
Batch process all your records using unstructured-ingest to store unstructured outputs locally on your filesystem and upload those to @clarifai apps.
Docs 👇
unstructured-io.github.io/unstructured/ingest/destination…
Clarifai
@clarifai
We are excited to introduce the Clarifai and @UnstructuredIO integration that allows you to streamline and optimize the data processing pipelines for LLMs. 🎉
The following notebook is a step-by-step guide on how to utilize Clarifai as your destination connector using @UnstructuredIO and import raw files from the S3 bucket. 👇
Checkout the Notebook here.
github.com/Clarifai/examples/blob/main/Integrations/Unstr… -
Clarifai
@clarifai
We are excited to introduce the Clarifai and @UnstructuredIO integration that allows you to streamline and optimize the data processing pipelines for LLMs. 🎉
The following notebook is a step-by-step guide on how to utilize Clarifai as your destination connector using @UnstructuredIO and import raw files from the S3 bucket. 👇
Checkout the Notebook here.
github.com/Clarifai/examples/blob/main/Integrations/Unstr… -
Clarifai
@clarifai
New release - Clarifai 10.3 is out now! ⚡️
🚀 We are introducing pre-built, ready-to-use templates that streamline the app creation process, making it faster and more efficient.
In addition to these templates, the release also includes great new features, new models, integrations, performance and UX improvements, and fixes.
Read the release blog here: www.clarifai.com/blog/clarifai-10.3-template-wizardry-cra… -
Clarifai
@clarifai
⭐ Clarifai is named a Leader in The Forrester Wave™: Computer Vision Tools, Q1 2024 ⭐
One of the key points that Forrester highlighted in their evaluation of Clarifai is the breadth of our best-of-breed Computer Vision capabilities and highly scalable and secure platform architecture that can benefit customers across industries and use cases.
Read the full report from Forrester here:
www.clarifai.com/report/the-forrester-wave-computer-visio… -
Clarifai
@clarifai
.@Cohere's latest LLM, Command R+, is now available on the Clarifai Platform! 🎉
• 104B parameters (open weights)
• 128k context length
• Multilingual coverage in 10 key languages
• Optimized for advanced Retrieval Augmented Generation (RAG) with citation to reduce hallucinations
Try it out here:
clarifai.com/cohere/generate/models/command-r-plus -
Clarifai reposted
SineWave Ventures
@SineWaveVC
Portfolio company @Clarifai was recognized by @Forrester as a leader in computer vision. The full-stack AI platform received the highest scores in multiple criteria. Read more: www.clarifai.com/report/the-forrester-wave-computer-visio…
-
Clarifai
@clarifai
Mistral Large from @MistralAI is available on the Clarifai Platform!
The model is:
- fluent in English, French, Spanish, German, and Italian
- 32K tokens context window
- instruction-following
- capable of function calling
Try it out: clarifai.com/mistralai/completion/models/mistral-large -
Clarifai
@clarifai
We are excited to be at @PyConLT 2024 🥳
If you are around don't miss the opportunity to hear from Isaac Chung, Senior Research Engineer from Clarifai, who will be delivering two exciting talks!
1. Transcend the Knowledge Barriers in RAG: Setup, Chat State, and more:
pycon.lt/2024/talks/HFXHRV
2. Speed up open source LLM-serving with llama-cpp-python:
pycon.lt/2024/talks/DHBLXW
Come join us! -
Clarifai reposted
Clarifai
@clarifai
A Gentle Introduction to Retrieval Augmented Generation (RAG) for the Intelligence Community!
LLMs are trained on large public datasets, but they don't have access to *your* data.
RAG has been revolutionizing how we access and utilize data, offering real-time information retrieval to enhance language models.
Deep dive into this blog where we explain the importance of RAG, especially for intelligence analysts with a real-world use case.
Read here: intelligencecommunitynews.com/ic-insiders-a-gentle-introd… -
Clarifai
@clarifai
A Gentle Introduction to Retrieval Augmented Generation (RAG) for the Intelligence Community!
LLMs are trained on large public datasets, but they don't have access to *your* data.
RAG has been revolutionizing how we access and utilize data, offering real-time information retrieval to enhance language models.
Deep dive into this blog where we explain the importance of RAG, especially for intelligence analysts with a real-world use case.
Read here: intelligencecommunitynews.com/ic-insiders-a-gentle-introd… -
Clarifai reposted
Clarifai
@clarifai
As we have announced our strategic partnership with @DeepgramAI, our goal was to integrate Deepgram's state-of-the-art speech-to-text technology expertise into the Clarifai full-stack AI platform, allowing developers to build AI faster with enhanced automatic speech recognition capabilities.
Tune in to the conversation featuring Scott Stephenson, CEO of Deepgram and Amith Nair, President and COO of Clarifai, to explore more about the partnership and how it empowers developers to build with AI faster. 🚀
Watch the full video here: youtu.be/hprwh2Lo8n8 -
Clarifai
@clarifai
As we have announced our strategic partnership with @DeepgramAI, our goal was to integrate Deepgram's state-of-the-art speech-to-text technology expertise into the Clarifai full-stack AI platform, allowing developers to build AI faster with enhanced automatic speech recognition capabilities.
Tune in to the conversation featuring Scott Stephenson, CEO of Deepgram and Amith Nair, President and COO of Clarifai, to explore more about the partnership and how it empowers developers to build with AI faster. 🚀
Watch the full video here: youtu.be/hprwh2Lo8n8 -
Clarifai
@clarifai
🎉 DBRX is available now for FREE on Clarifai! 🎉
We're excited to announce some big news heading into the weekend! @clarifai is now offering @databricks DBRX for free to all our community customers! #dbrx
www.linkedin.com/feed/update/urn:li:activity:717951357034… -
Clarifai
@clarifai
We are thrilled to announce that Clarifai has been named a Leader in The Forrester Wave™: Computer Vision Tools, Q1 2024.
Read the announcement here:
www.clarifai.com/press-release/clarifai-named-a-leader-fo… -
Clarifai reposted
Clarifai
@clarifai
Evaluate your fine-tuned LLMs. 🚀
A powerful approach to improving the applicability of LLMs to specific tasks is through fine-tuning.
Once you've fine-tuned LLMs on specific data, it's essential to evaluate their performance and effectiveness in real-world scenarios.
You can now easily fine-tune LLMs using the Clarifai Platform. After fine-tuning, you can also assess their performance using the newly introduced LLM Evaluation module.
The module allows you to evaluate LLMs against standardized benchmarks and custom criteria.
This guide covers a step-by-step process to fine-tune and evaluate your LLMs:
Read more:
docs.clarifai.com/portal-guide/evaluate/llms/ -
Clarifai
@clarifai
Evaluate your fine-tuned LLMs. 🚀
A powerful approach to improving the applicability of LLMs to specific tasks is through fine-tuning.
Once you've fine-tuned LLMs on specific data, it's essential to evaluate their performance and effectiveness in real-world scenarios.
You can now easily fine-tune LLMs using the Clarifai Platform. After fine-tuning, you can also assess their performance using the newly introduced LLM Evaluation module.
The module allows you to evaluate LLMs against standardized benchmarks and custom criteria.
This guide covers a step-by-step process to fine-tune and evaluate your LLMs:
Read more:
docs.clarifai.com/portal-guide/evaluate/llms/ -
Clarifai reposted
Clarifai
@clarifai
An instruction-generation model designed to create valid instructions given a raw text corpus. 📝
Introducing Genstruct-7B, an LLM from @NousResearch that enables the creation of new, partially synthetic instruction fine-tuning datasets from any raw-text corpus. 🚀
The model focuses on generating complex questions and multi-step reasoning for each instruction pair, surpassing previous models that only generate direct questions and responses.
It offers superior quality and diversity in the generated data.
Check out the model, try it with a few inputs, and also access it via an API using the Python SDK: 👇
clarifai.com/nousresearch/instruction-generation/models/g… -
Clarifai
@clarifai
An instruction-generation model designed to create valid instructions given a raw text corpus. 📝
Introducing Genstruct-7B, an LLM from @NousResearch that enables the creation of new, partially synthetic instruction fine-tuning datasets from any raw-text corpus. 🚀
The model focuses on generating complex questions and multi-step reasoning for each instruction pair, surpassing previous models that only generate direct questions and responses.
It offers superior quality and diversity in the generated data.
Check out the model, try it with a few inputs, and also access it via an API using the Python SDK: 👇
clarifai.com/nousresearch/instruction-generation/models/g… -
Clarifai
@clarifai
RAG with PDFs
Want to easily set up an RAG pipeline for PDFs?
Check out our blog on how to build it exactly using @langchain.
Blog: clarifai.com/blog/ai-in-5-rag-with-pdfs -
Clarifai
@clarifai
The Mxbai-embed-large-v1 is a SOTA sentence embedding model designed to capture the nuanced semantics of text across a wide range of domains and tasks!
The model is trained using AnglE loss, a technique that ensures the model learns high-quality, context-rich embeddings.
Mxbai-embed-large-v1 outperforms the new OpenAI embeddings model, text-embedding-3-large, and also matches the performance of 20x larger models like echo-mistral-7b.
You can now run the Mxbai-embed-large-v1 using Clarifai's Python SDK with just a few lines of code.
Try it out: clarifai.com/mixedbread-ai/embed/models/mxbai-embed-large-v1 -
Clarifai
@clarifai
What are your thoughts on Nvidia's announcement this week in #gtc24 ?
@_chriskent shares his perspective on the groundbreaking announcement, and why the future of AI extends beyond hardware and will require flexibility and choice to meet diverse user needs.
Read the article here:
www.clarifai.com/blog/nvidias-breakthrough-and-hybrid-ai-… -
Clarifai
@clarifai
.@DeepgramAI introduced Aura: Lightning Fast Text-to-Speech model built for responsive, conversational AI agents and applications.
Here are some key features of Aura:
✅ Lightning-fast speed with less than 250 ms latency
✅ High-quality voices with natural-sounding tone, rhythm, and emotion
✅ Cost-efficient for high-throughput applications
You can now access Aura within the Clarifai Platform and run the model using Clarifai’s Python SDK in just a few lines of code.
Try it out here: 👇
clarifai.com/deepgram/tts/models/aura-tts
#texttospeech -
Clarifai
@clarifai
Last Chance Reminder: Join us today for the AI Mixer at #gtc24!
If you are heading to Nvidia GTC, don't miss out on our exclusive AI Mixer with Clarifai at the Continental Bar today, March 19th, from 6-10:30 PM.
Connect with AI leaders and builders, enjoy drinks and food, and exchange ideas from the conference and the future of AI!
RSVP here: lu.ma/ai-mixer-sanjose -
Clarifai
@clarifai
Join us tomorrow for the AI Mixer at #gtc24! ⚡️
Heading to Nvidia GTC? Come join us tomorrow for an exclusive AI Mixer with Clarifai at the Continental Bar on March 19th from 6-10:30 PM.
Connect with AI leaders and builders, enjoy drinks and food, and exchange ideas from the conference and the future of AI!
RSVP here: lu.ma/ai-mixer-sanjose -
Clarifai
@clarifai
Clarifai is thrilled to announce our strategic partnership with @DeepgramAI, a leader in automatic speech recognition (ASR) technology. 🚀
Our goal was to integrate Deepgram's state-of-the-art speech-to-text technology expertise into the Clarifai full stack AI platform, allowing developers to build AI faster with enhanced ASR capabilities.
Read the announcement to learn more about the key highlights of the partnership:
www.clarifai.com/press-release/clarifai-and-deepgram-anno… -
Clarifai
@clarifai
Only 4 days left for the AI Mixer at NVIDIA GTC! 🔥
Heading to Nvidia GTC? come join us for an exclusive AI Mixer with Clarifai at the Continental Bar on March 19th from 6-10:30 PM.
Connect with AI leaders and builders, enjoy some drinks and food, and exchange ideas from the conference and the future of AI!
Secure your spot here: lu.ma/ai-mixer-sanjose
*Space is limited, but if you wish to bring a guest, just let us know, and we can do our best to accommodate. -
Clarifai
@clarifai
Today, we're excited to announce the launch of the AI Developer Alliance, a coalition of developer-led organizations dedicated to shaping the future of responsible and groundbreaking AI advancement. ⚡️
Our charter is built upon three core principles:
• Advocacy: We champion responsible, ethical, and inclusive AI practices and policies that prioritize transparency and accountability.
• Expertise: We bring together diverse experts from fields such as computer science, engineering, and mathematics to promote expertise and support within the AI community.
• Trust: We cultivate an open and collaborative community advocating for transparent, ethical AI development that fosters trust and drives responsible innovation.
By uniting diverse minds and perspectives, our goal is to accelerate learning, innovation, and responsible AI practices globally. 🌍
Does this resonate with you? Join the conversation with like-minded industry experts and help shape the future of AI.
developeralliance.ai -
Clarifai
@clarifai
As organizations deploy AI at scale, the community has a responsibility to guide global AI advancement, best practices, and ethics!
Introducing the AI Developer Alliance, a cohort of developer-led organizations who will come together to push the boundaries of AI.
Read the announcement here:
www.clarifai.com/press-release/clarifai-announces-ai-deve… -
Clarifai
@clarifai
6 days to go for the AI Mixer at NVIDIA GTC! 🚀
If you are going to Nvidia GTC, come join us for an exclusive AI Mixer with Clarifai at the Continental Bar on March 19th from 6-10:30 PM.
Connect with AI leaders and builders, enjoy some drinks and food, and exchange ideas from the conference and the future of AI!
RSVP here: lu.ma/ai-mixer-sanjose
*Space is limited, but if you wish to bring a guest, just let us know, and we can do our best to accommodate. -
Clarifai
@clarifai
New release - Clarifai 10.2 is out now! ⚡️
🚀 With a set of new great features, performance and UX improvements, and fixes
Read the blog post here:👇
www.clarifai.com/blog/clarifai-10.2-report-card-for-your-… -
Clarifai
@clarifai
Going to Nvidia GTC? 🚀
Come join us for an exclusive AI Mixer with Clarifai at the Continental Bar on March 19th from 6-10:30PM.
Connect with AI leaders and builders and enjoy some drinks and food, and exchange ideas from the conference and the future of AI!
If you're interested, you can RSVP here: lu.ma/ai-mixer-sanjose
*Space is limited, but if you wish to bring a guest, just let us know, and we can do our best to accommodate! -
Clarifai
@clarifai
Claude 3 comes with vision support similar to the GPT-4 Vision model!
Here is how you can pass your image inputs along with the prompts in Python with just a few lines of code ⚡️
Check out the code and try out the model here:
clarifai.com/anthropic/completion/models/claude-3-opus -
Clarifai reposted
-
Clarifai
@clarifai
Introducing 1-bit LLMs! ⚡️
This represents a new paradigm for LLM quantization, where every single parameter is ternary {-1, 0, 1} instead of 16-bit.
It makes the models 2.7x faster, uses 3.5x less GPU memory, and is 71x more energy efficient.
arxiv.org/abs/2402.17764v1 -
Clarifai
@clarifai
Retrieval Augmented Generation (RAG) in 4 lines of code! ⚡ 🚀
Read our latest blog on how to build a RAG system in 4 lines of code in Python.
www.clarifai.com/blog/retrieval-augmented-generation-rag-… -
Clarifai reposted
Clarifai
@clarifai
.@AnthropicAI released Claude 3, outperforming the GPT-4 model from OpenAI! ⚡️
Here are some key highlights: 🔥
- Three different models: Haiku, Sonnet, Opus.
- Opus model outperforms GPT-4 on every benchmark.
- 200K token context window.
- Sonnet is 2x faster than the predecessors
- Opus reduces wrong responses and doubles the accuracy.
- Comes with Vision support similar to GPT-4 Vision model
Try out Claude 3 now and access it via an API using Clarifai Python SDK in just a few lines of code here: 👇
Claude 3 Opus: clarifai.com/anthropic/completion/models/claude-3-opus
Claude 3 Sonnet: clarifai.com/anthropic/completion/models/claude-3-sonnet -
Clarifai
@clarifai
.@AnthropicAI released Claude 3, outperforming the GPT-4 model from OpenAI! ⚡️
Here are some key highlights: 🔥
- Three different models: Haiku, Sonnet, Opus.
- Opus model outperforms GPT-4 on every benchmark.
- 200K token context window.
- Sonnet is 2x faster than the predecessors
- Opus reduces wrong responses and doubles the accuracy.
- Comes with Vision support similar to GPT-4 Vision model
Try out Claude 3 now and access it via an API using Clarifai Python SDK in just a few lines of code here: 👇
Claude 3 Opus: clarifai.com/anthropic/completion/models/claude-3-opus
Claude 3 Sonnet: clarifai.com/anthropic/completion/models/claude-3-sonnet -
Clarifai
@clarifai
Join us today at 10:30 am ET for the live session on building a simple RAG system with DSPy!
Event Link: 👇
discord.gg/8Udh6uEFEB?event=1211717743372410961 -
Clarifai reposted
Clarifai
@clarifai
DSPy is the new framework that lets you program language models, not just prompting!
In this session, we will walk you through what DSPy is and how you can build a simple RAG system with DSPy and Clarifai! ⚡️
Join us tomorrow at 10:30 am ET: 👇
discord.gg/8Udh6uEFEB?event=1211717743372410961 -
Clarifai
@clarifai
DSPy is the new framework that lets you program language models, not just prompting!
In this session, we will walk you through what DSPy is and how you can build a simple RAG system with DSPy and Clarifai! ⚡️
Join us tomorrow at 10:30 am ET: 👇
discord.gg/8Udh6uEFEB?event=1211717743372410961 -
Clarifai
@clarifai
Join us in welcoming Chris Kent to clarifai as our new SVP of Marketing, set to propel growth as enterprises seek to build AI faster with Clarifai. An exciting addition to our team! www.clarifai.com/press-release/clarifai-hires-chris-kent-…
-
Clarifai reposted
Clarifai
@clarifai
Google released Gemma 2B and 7B a new family of lightweight and state-of-the-art open-source language models!
You can now run these models with an API in just a few lines of code:
Gemma 7B: clarifai.com/gcp/generate/models/gemma-7b-it -
Clarifai
@clarifai
Google released Gemma 2B and 7B a new family of lightweight and state-of-the-art open-source language models!
You can now run these models with an API in just a few lines of code:
Gemma 7B: clarifai.com/gcp/generate/models/gemma-7b-it
Clarifai
@clarifai
-
Clarifai
@clarifai
DSPy is the framework that lets you program language models, not just prompting! 🔥
Our latest tutorial will walk you through DSPy and building a simple RAG system with DSPy and Clarifai! ⚡️
Watch the full video here: youtu.be/NxFj8Vv0fWY -
Clarifai
@clarifai
Attending the CDAO Advantage DoD 2024 - Defense Data & AI Symposium in DC today? Stop by the Clarifai table #87 to learn how to build secure and trust-worthy AI applications to boost your organization.
Connect with our onsite team at table 87. www.clarifai.com/events/advantage-dod-2024?hss_channel=tw…
#adod24 -
Clarifai
@clarifai
Welcome Amith Nair to @clarifai as our new President and COO, ready to take on today's challenges of helping enterprises operationalize AI at scale while optimizing for speed, energy, and cost. www.clarifai.com/press-release/clarifai-hires-amith-nair-…
-
Clarifai
@clarifai
Are you going to be at Advantage DoD 2024?
Visit us at table 87 to meet our AI experts, explore live demonstrations, and learn how Clarifai can elevate your generative AI capabilities. www.clarifai.com/events/advantage-dod-2024?hss_channel=tw…
#adod24 #generativeai #ai #defense #government -
Clarifai
@clarifai
Advantage DoD 2024 starts tomorrow at Washington, DC. Visit us at booth 87 to learn how you can access a wide array of AI tools to build reliable and trustworthy AI for your DoD.
See you tomorrow! www.clarifai.com/events/advantage-dod-2024?hss_channel=tw…
#generativeai #defense #dod #government #ai #adod24 -
Clarifai
@clarifai
Clarifai is going to Advantage DoD 2024 in Washington, DC. from Feb 20-22!
Visit us at table 87 to learn how you can empower your DoD with Clarifai's full stack AI platform.
See you there. www.clarifai.com/events/advantage-dod-2024?hss_channel=tw…
#ai #generativeai #dod #government -
Clarifai
@clarifai
Join us today at 10:30 am ET for the live session on building a RAG system in just 4 lines of code in Python!
Event Link: 👇
discord.gg/BYZYv3xF?event=1207229498937049128 -
Clarifai
@clarifai
Build a RAG system in 4 lines of code! 🔥
Join us tomorrow at 10:30 am ET for the live session, where we will work through building a RAG system in just 4 lines of code in Python using Clarifai's Python SDK!
Join here: 👇
discord.gg/BYZYv3xF?event=1207229498937049128 -
Clarifai
@clarifai
Join us to learn from the Winners of the NextGen GPT AI Hackathon
starting in less than an hour! 📣
Event Link: ➡️ discord.gg/6Bu7DDZe?event=1204461908770684968 -
Clarifai
@clarifai
Join the Project Showcase featuring the winners of Clarifai's NextGen GPT AI Hackathon!🔥
Event Link: ➡️ discord.gg/6Bu7DDZe?event=1204461908770684968
The Top 3 teams: Story Prep, AI Project Manager Assistant and Recovery Pal will present their ideas and their approach to building these projects! -
Clarifai reposted
Clarifai
@clarifai
Here are the winners in our recent NextGen GPT AI Hackathon with Clarifai and @lablabai !🎉
1st Place 🥇: Story Prep
An app that helps you learn a language by reading your favourite author. All you need to do is upload your book/story.
Read more: 🧵 ⬇️
lablab.ai/event/nextgen-gpt-ai-hackathon/story-prep/story… -
Clarifai
@clarifai
Here are the winners in our recent NextGen GPT AI Hackathon with Clarifai and @lablabai !🎉
1st Place 🥇: Story Prep
An app that helps you learn a language by reading your favourite author. All you need to do is upload your book/story.
Read more: 🧵 ⬇️
lablab.ai/event/nextgen-gpt-ai-hackathon/story-prep/story…
Clarifai
@clarifai
This will convert the story into lessons that teach just the necessary for you to enjoy that particular story.
lablab.ai/event/nextgen-gpt-ai-hackathon/story-prep/story…
Clarifai
@clarifai
2nd Place 🥈: AI Project Manager Assistant
AI Project Manager Assistant is a Slack bot that automatically records meeting transcripts and sends the summary in a specified Slack channel.
lablab.ai/event/nextgen-gpt-ai-hackathon/jarvis/ai-projec…
Clarifai
@clarifai
3rd Place 🥉: Recovery Pal
Recovery Pal is a mobile app designed for addiction recovery.
The app supports features such as Emotional Check-in, Custom Affirmations, Journaling for self-reflection, Art Interpretation, and finally custom meditations:
lablab.ai/event/nextgen-gpt-ai-hackathon/recoverypal/reco… -
Clarifai
@clarifai
We're ready!
See you all in a few mins as we begin our Hands-on Workshop on How To Build A Model at Carahcafe in Reston, VA.
We will be learning how to leverage the power of AI to solve critical challenges. -
Clarifai
@clarifai
Introducing Code Llama 70B: a new, more performant LLM from @AIatMeta for code generation!
Try and experiment with Code Llama 70B for free in the @clarifai community! 😍
👉 clarifai.com/meta/Llama-2/models/codeLlama-70b-Instruct
You can also access and run the model with just a few lines of code:
Read more:
Clarifai
@clarifai
With the Clarifai Python SDK you can access this model with just a few lines of code:
Check out the code in the model notes here:
clarifai.com/meta/Llama-2/models/codeLlama-70b-Instruct
Clarifai
@clarifai
We also have the Code Llama 70B Python which is specialized in Python code generation
Try it now here: 👇
clarifai.com/meta/Llama-2/models/codeLlama-70b-Python -
Clarifai
@clarifai
Excited for tomorrow!
The Clarifai team will be in Reston, Virginia on Jan 31 to host a hands-on workshop on How To Build An AI Model. Join us to learn how to leverage the power of AI and the Clarifai platform in your mission-critical tasks. Register now! carahevents.carahsoft.com/Event/Details/436823-Clarifai?h… -
Clarifai
@clarifai
Spend an afternoon with Clarifai to build an AI model and learn how to apply AI in your daily work. Join our hands-on workshop in person in Reston, VA, on Jan 31. Register for free carahevents.carahsoft.com/Event/Details/436823-Clarifai?h…
-
Clarifai
@clarifai
AI can reduce the risks and errors in mission-critical projects. With Clarifai you can build AI seamlessly with all the tools under one platform. Join us in Reston, VA, on Jan 31 to learn how to build a model with the Clarifai platform.
Register for free carahevents.carahsoft.com/Event/Details/436823-Clarifai?h… -
Clarifai
@clarifai
AI can drastically improve the success of your high-impact projects. Its time for organizations to harness the power of AI to solve mission-critical challenges. Join our hands-on workshop in Reston, VA on Jan 31 to learn how to build a model easily. carahevents.carahsoft.com/Event/Details/436823-Clarifai?h…
-
Clarifai
@clarifai
We are hosting a live technical workshop starting in 2 hours with @lablabai for the NextGen GPT AI hackathon participants!🔥
Join us today at 9:30 am ET to learn about RAG and how to build RAG systems:
Live stream link: www.twitch.tv/lablabai -
Clarifai
@clarifai
Mixtral 8x7B paper was released a few days back!
Mixtral 8x7B, a Sparse Mixture of Experts (SMoE) language model has same architecture as Mistral 7B, with the difference that each layer is composed of 8 feedforward blocks called as experts
Paper: arxiv.org/abs/2401.04088 -
Clarifai reposted
Afflatus
@useafflatus
"AI In 5: RAG With PDFs", an article by @clarifai which combines their technology with @langchain to enable RAG 🧵
-
Clarifai reposted
lablab.ai
@lablabai
🌐Dive into AI's Latest with Clarifai's Models!
👨🏫 Join @itspauloalmeida , Mentor Lead at lablab.ai, for a walkthrough of @clarifai 's newest AI technologies, including GPT-4 Turbo, Image Gen, and Voice Recognition.
👉 This guide is perfect for anyone looking to stay ahead in AI advancements and be ready for the Hackathon that starts this Friday January 12th!
📺 Watch full video here: www.youtube.com/watch?v=ECIMoNilC2I&t=4s
🚀 Enhance your skills at the NextGen GPT AI Hackathon. Details here: lablab.ai/event/nextgen-gpt-ai-hackathon -
Clarifai
@clarifai
3 hours to go for the kickoff 🥳
Join us for the NextGen GPT AI Hackathon kickoff stream today at 12pm ET as we will be introducing you to the Clarifai to build Generative AI Apps!
If you haven't registered yet for the hackathon?
Register here: lablab.ai/event/nextgen-gpt-ai-hackathon -
Clarifai
@clarifai
3 days to go for the @Clarifai’s NextGen GPT AI Hackathon with
@Lablabai ! 📣
Join us to build the best AI-powered project with GPT-4 Vision and other cool AI models!
A total of $4,500 in prizes up for grabs! 💰
Register here: lablab.ai/event/nextgen-gpt-ai-hackathon -
Clarifai
@clarifai
Join us today for the amazing X Space!
We'll be there with the @lablabai team, discussing various AI models such as GPT-4 Turbo, Image Gen, Voice & Custom Models, and the state of AI!
x.com/i/spaces/1BdGYrboRqBJX -
Clarifai
@clarifai
We're thrilled to bring you @Clarifai’s NextGen GPT AI Hackathon with @Lablab.ai! 📣
A total of $4,500 in prizes up for grabs! 💰
Join us and build your best AI-powered project with GPT-4 Vision and other cool models!
Register here: lablab.ai/event/nextgen-gpt-ai-hackathon -
Clarifai
@clarifai
A Free Machine Learning course from Stanford taught by Andrew Ng.
You'll learn:
• Linear/Logistic regression
• Neural Networks
• Decision Trees
• Support Vector Machines
• Naive Bayes
• KNN
• Gradient Descent
And much more.
youtube.com/playlist?list=PLoROMvodv4rMiGQp3WXShtMGgzqpfVfbU -
Clarifai
@clarifai
GPT using Numpy! 🔥
Here is Generative Pretrained Transformer(GPT) implemented from scratch using Numpy in just 60 lines of code:
Clarifai
@clarifai
Although the above code is super slow and doesn't include any training code, backpropagation, GPU/TPU support, batching, etc
This is a good way to get an idea of the implementation of GPT.
Checkout the complete code here:
github.com/jaymody/picoGPT -
Clarifai
@clarifai
📚 Clarifai acknowledged for three Use Cases in the 2023 Gartner® Critical Capabilities for Cloud AI Developer Services report. With our introduction of Large Language Models and our role in shaping the future of language AI, we're at the forefront.
www.clarifai.com/report/gartner-cloud-ai-developer-servic… -
Clarifai
@clarifai
Mathematics for Machine Learning from University of Berkeley.
Learn:
• Linear Algebra
• Calculus and Optimization
• Probabilities
100% Free.
Download it here: gwthomas.github.io/docs/math4ml.pdf -
Clarifai reposted
LangChain
@LangChain
🌲RAG With PDFs
Want to easily set up a RAG pipeline over PDFs?
This great tutorial by our friends at @clarifai goes over how to do exactly that
Blog: www.clarifai.com/blog/ai-in-5-rag-with-pdfs -
Clarifai
@clarifai
Clarifai is included in @my_hatchpad's latest article on innovative tech start-ups to work for. www.hatchit.io/15-innovative-startups-on-the-hunt-for-tec…
Want to work at Clarifai? Check our job openings - www.clarifai.com/company/careers?hss_channel=tw-2401933753 -
Clarifai reposted
Clarifai
@clarifai
Join us to learn about Retrieval Augmented Generation (RAG) as we work through building a generative question answering system using RAG from start to finish. 📌
Date: 21st December | 10:30 am ET
discord.gg/NRpxuPvn?event=1186525179539640401 -
Clarifai
@clarifai
🚀 Clarifai recognized for Vision, Language, and autoML Use Cases in the latest 2023 Gartner® Critical Capabilities for Cloud AI Developer Services. Explore the "Critical Capabilities for Cloud AI Developer Services" report for a closer look.
www.clarifai.com/report/gartner-cloud-ai-developer-servic… -
Clarifai
@clarifai
Mixtral 8x7B is a high-quality Sparse Mixture-of-Experts model from @MistralAI
Mixtral outperforms Llama 2 70B on most benchmarks with 6x faster inference.
clarifai.com/mistralai/completion/models/mixtral-8x7B-Ins…
Access the Mixtral model now with an API: 🧵👇Post not available here, view on X
Clarifai
@clarifai
As mentioned mixtral outperforms Llama 2 70B, as well as GPT3.5, on most benchmarks.
Clarifai
@clarifai
Mixtral 8x7B Instruct also released alongside Mixtral 8x7B.
This model has been optimised through supervised fine-tuning for careful instruction following.
Try out the model now and access it with an API: 👇
clarifai.com/mistralai/completion/models/mixtral-8x7B-Ins… -
Clarifai
@clarifai
Join us to learn about Retrieval Augmented Generation (RAG) as we work through building a generative question answering system using RAG from start to finish. 📌
Date: 21st December | 10:30 am ET
discord.gg/NRpxuPvn?event=1186525179539640401 -
Clarifai
@clarifai
If you want to get started with machine learning and want a hands-on tutorial:
Intro to Machine Learning from Kaggle will help you build your first model!
• 8 Lessons.
• 100% Free.
Check this out:
kaggle.com/learn/intro-to-machine-learning -
Clarifai
@clarifai
Deep Learning course from Carnegie Mellon University.
One of the best you'll find out there to learn Deep Learning fundamentals.
100% FREE
Check this out:
youtube.com/playlist?list=PLp-0K3kfddPxRmjgjm0P1WT6H-gTqE8j9 -
Clarifai
@clarifai
Build a RAG system with @langchain and Clarifai!
Here is a step by step walkthrough of building Retrieval Augmented System for Generative Question Answering.
Check this out: youtu.be/4KkZM8JVn-0 -
Clarifai
@clarifai
Join us for the live session today at 10:30 am ET on Cross-modal labeling and Transfer Learning.
Discord Stage: discord.gg/gbfBsj3T?event=1183824601281740821 -
Clarifai
@clarifai
Introducing Llama-Guard, a 7B parameter input-output safeguard model that excels in classifying safety risks in Human-AI conversations! 🔥
clarifai.com/meta/Llama-2/models/llamaGuard-7b
You can now run LlamaGuard 7B with an API: 🧵
Clarifai
@clarifai
LlamaGuard is built on top of Llama2-7b, and utilizes a safety risk taxonomy, distinguishing between safety risks in user prompts (prompt classification) and generated responses (response classification).
Check out how LlamaGuard 7B compared with other models!
Clarifai
@clarifai
Want to try the Llama Guard 7B model and access it with an API : 🚀
Check this out here: clarifai.com/meta/Llama-2/models/llamaGuard-7b -
Clarifai
@clarifai
Introducing Cybertron 7B v2, a 7B MistralAI-based model, the best in its series.
The model excels in mathematics, logic, and reasoning!
clarifai.com/fblgit/una-cybertron/models/una-cybertron-7b-v2
Read more ⬇️
Clarifai
@clarifai
The model trained using SFT, DPO, and UNA (Unified Neural Alignment) techniques on diverse datasets.
Checkout the leaderboard as of December 5th here:
Clarifai
@clarifai
Want to test the model?
We've added it to the platform. Try it out here:
clarifai.com/fblgit/una-cybertron/models/una-cybertron-7b-v2 -
Post not available here, view on X
Clarifai
@clarifai
The model trained using SFT, DPO, and UNA (Unified Neural Alignment) techniques on diverse datasets.
Checkout the leaderboard as of December 5th here:
Clarifai
@clarifai
Want to test the model?
We've added it to the platform. Try it out here:
clarifai.com/fblgit/una-cybertron/models/una-cybertron-7b-v2 -
Clarifai
@clarifai
🖼️ Clarifai recognized for Vision, Language, and autoML Use Cases in the latest 2023 Gartner® Critical Capabilities for Cloud AI Developer Services. Explore the "Critical Capabilities for Cloud AI Developer Services" report for a closer look.
www.clarifai.com/report/gartner-cloud-ai-developer-servic… -
Clarifai
@clarifai
Join us for a live session on Cross-modal labeling and Tansfer Learning!
discord.gg/gbfBsj3T?event=1183824601281740821
Learn how to leverage cross modal labeling, to do transfer learning on multiple models and evaluate them effectively.
When: Dec 14th at 10:30 am ET -
Clarifai
@clarifai
Building applications with GPT-4 Vision!
The paper discusses emerging application scenarios and future research directions for GPT-4V based systems.
Check this out: arxiv.org/pdf/2309.17421.pdf -
Clarifai
@clarifai
Arguably, the best Linear Algebra course out there for Machine Learning.
Taught by MIT's Professor Gilbert Strang.
And it's FREE!
Build a strong foundation in mathematics for machine learning:
youtube.com/playlist?list=PL221E2BBF13BECF6C -
Clarifai
@clarifai
Cheatsheet and a quick reference guide that covers:
- LLM basics
- Word embeddings and vector representations
- Transformer, GPT etc architecture
- Training and fine-tuning LLMs
- Evaluation metrics for LLMs
Check this out:
github.com/Abonia1/CheatSheet-LLM -
Clarifai
@clarifai
Glad to be included in @my_hatchpad's latest article on innovative tech start-ups to work for. www.hatchit.io/15-innovative-startups-on-the-hunt-for-tec…
Want to work at Clarifai? Check our job openings - www.clarifai.com/company/careers?hss_channel=tw-2401933753 -
Clarifai
@clarifai
Mathematics for Machine Learning from University of Berkeley!
100% Free.
Learn:
• Linear Algebra
• Calculus and Optimization
• Probabilities
Download it here: gwthomas.github.io/docs/math4ml.pdf -
Clarifai
@clarifai
Productionizing Generative AI for Government 🔥
As government agencies prepare to adopt generative AI, are you mission-ready for 2024?
clarifai.com/webinar/productionizing-generative-ai-for-go…
Join Clarifai CEO, Matt Zeiler Ph.D., and VP of Public Sector Jonathan Padgett, as they present solutions to AI adoption from a technical and AI infrastructure perspective. -
Clarifai
@clarifai
Is your organization prepared to adopt and deploy Generative AI?
Join us for a webinar on December 6 to hear how to craft a winning Generative AI strategy for 2024.
👉 Register here: www.clarifai.com/webinar/productionizing-generative-ai-fo…
📆 Date: December 6th 2023
🕑 Time: 1:30 pm ET | 10:30 am PT -
Clarifai
@clarifai
Productionizing AI for Government!
Building GenAI on top of a proven full stack AI platform is the only way to go.
Join CEO Matt Zeiler, Ph.D. and Jonathan Padgett, VP Public Sector to learn how Clarifai does this and can help you today!
Register here:
www.clarifai.com/webinar/productionizing-generative-ai-fo… -
Clarifai reposted
LDV Capital
@ldvcapital
Building GenAI on top of a proven full stack AI platform is the only way to go. Join CEO @mattzeiler, Ph.D. and Jonathan Padgett, VP Public Sector to learn how @clarifai does this and can help you today! Productionizing AI for Government, register here! www.clarifai.com/webinar/productionizing-generative-ai-fo…
-
Clarifai
@clarifai
LLMs for Generating Structured Data! 🔥
This paper assess the capability of current LLMs in generating complex structured data and also proposed a structure-aware fine-tuning approach as a solution to improve this ability of LLMs
The authors of the paper applied the structure-aware fine-tuning method, to LLaMA-7B, which significantly outperformed other evaluated LLMs.
Check this out:
arxiv.org/abs/2309.08963 -
Clarifai
@clarifai
Join us today at 10:30 am ET for a live session on how to do transfer learning on Large Language Models!
Check this out: ⬇️
discord.gg/cxpyNM5D?event=1177254020402913412 -
Clarifai
@clarifai
Join us tomorrow for a live session on Transfer Learning LLMs!
discord.gg/cxpyNM5D?event=1177254020402913412
This is a step-by-step guide on transfer learning a large language model, specifically the Cohere Embedding Model for Text Classification! 28th
28th Nov: 10:30 am ET -
Clarifai
@clarifai
Applied Machine Learning from Cornell Tech!
Learn:
- Supervised and Unsupervised Machine Learning Techniques
- Deeplearning & Generative Models
- Error Analysis, Baselines and Evaluations.
youtube.com/playlist?list=PL2UML_KCiC0UlY7iCQDSiGDMovaupqc83 -
Clarifai
@clarifai
Let's discuss your generative AI needs at re:invent! Clarifai's full stack platform will help you build and deploy mission-critical AI applications at scale.
Schedule a few minutes with me:
www.clarifai.com/events/reinventmatt-zeiler
#awsreinvent #reinvent2023 #fullstackaiplatform #clarifai -
Clarifai
@clarifai
Join us for a session on Transfer Learning with LLMs! 🔥
Get a step-by-step guide on using a large language model, specifically the Cohere Embedding Model, to build a text classification model.
⏰ 28th Nov: 10:30 am ET
discord.gg/2hAADNUW?event=1177254020402913412 -
Clarifai
@clarifai
Access the latest Claude 2.1 from @AnthropicAI with an API!
1. Claude 2.1 provides 200K token context window, doubling the amount of information it can process.
2. 2x Decrease in Hallucination Rates
clarifai.com/anthropic/completion/models/claude-2_1
Read more: ⬇️
Clarifai
@clarifai
Read the Claude 2.1 model notes here:
www.clarifai.com/blog/run-claude-2.1-with-an-api -
Clarifai
@clarifai
Advanced NLP from Carnegie Mellon University!
The course covers current state-of-the-art NLP techniques & algorithms.
One of the best NLP courses on the web!
Check this out:
youtube.com/playlist?list=PL8PYTP1V4I8D0UkqW2fEhgLrnlDW9QK7z -
Clarifai
@clarifai
Applications of Transformers!
A comprehensive survey paper on applications of Transformers for deep learning tasks.
This also includes list of transformer models for NLP, Computer Vision, Multi-Modal and Audio.
arxiv.org/abs/2306.07303 -
Clarifai
@clarifai
Clarifai turned 10 🎉
To Celebrate this moment, on our 10 year anniversary we are announcing the Full Stack AI Platform.
The platform offers end-to-end capabilities, including serverless inference, prompt engineering, model comparison and routing, low-rank fine-tuning, evaluations, serverless vectorDB, auto-indexing, generative data labeling, and model monitoring.
Now, building generative AI use cases such as retrieval augmented generation, document AI, visual generation, multi-modal search, and more, seamlessly from prototype to production gets easier.
Check out our latest official press release on Full Stack AI Platform. Get a glimpse of the AI innovations with Clarifai platform!
www.clarifai.com/press-release/a-decade-of-building-the-f…
Also to celebrate the big 10, we're giving away 1 hour/month of deep fine-tuning to all Clarifai users forever so you can produce accurate insights and predictions. 🎁
The offer is auto-applied for each user so simply log into the Clarifai platform to start using your free hour of deep training.
Follow these steps to deep fine-tune your models: 👇
docs.clarifai.com/portal-guide/model/deep-training/ -
Clarifai
@clarifai
Introducing AI Lake! 🔥
One central place to share and reuse all the AI models, configurations, annotations, datasets, workflows, UI Modules, and more!
AI Lake let's you:
• Produce and reproduce AI at scale
• Increase the value of your data and human capital
• Reduce AI expenses
• Secure and protect your AI
Read more here: 👇
www.clarifai.com/solution-brief/clarifai-ai-lake -
Clarifai
@clarifai
Build RAG Pipelines with @langchain and Clarifai!
LLMs have the drawback of being too generic. They lack domain-specific knowledge and they hallucinate a lot as they are not up to date.
Retrieval Augmented Generation (RAG), connects LLMs to an external database to fetch up-to-date or context-specific data and make it available to an LLM when asking it to generate a response.
You can now build RAG systems with Clarifai:
Check out the Notebook here: 👇
colab.research.google.com/drive/1NldVhlYnjyg1A0KEoxnr5SmG… -
Clarifai
@clarifai
3 Usecases and Benefits of Vector Databases 🔥
1. Retrieval Augmented Generation:
RAG process is particularly useful in situations where being up-to-date is necessary. It empowers the LLM to fetch and use the most recent and relevant information beyond its original training data. Compared to building your own foundation model or fine-tuning an existing model for context-specific issues, RAG is cost-effective and easier to implement.
2. Semantic Search:
Semantic search uses vectors to search and retrieve text, images and videos. Compared to traditional keyword search, vector search yields more relevant results and executes faster.
3. Multimodal Search:
Multimodal search is a specific case of semantic search. It represents a paradigm shift from traditional search methods, allowing users to query across diverse data types, such as text, images, audio, and video.
Read more here: 👇
www.clarifai.com/blog/use-cases-and-benefits-of-vector-da… -
Clarifai
@clarifai
Import models from @huggingface to the Clarifai Platform in 4 simple steps.
Step 1: Assess the model's compatibility for importing into the Clarifai platform.
Step 2: Go to your Clarifai App and then, click Add a model in the model section and select Import from Hugging Face.
Step 3: Pass the model ID, model notes, Hugging Face pipeline name, model name and import.
Step 4: You can now access this model in your Clarifai Application with the Use Model option.
Check this out for a detailed step by step guide: 👇
docs.clarifai.com/portal-guide/model/hf-model-importer/ -
Clarifai
@clarifai
Build an AI App in Python! 🔥
With the new Python SDK, the functionality of the Clarifai platform is now accessible with fewer lines of code!
Learn how you can build an AI App in Python with the Python SDK.
Check this out: 👇
www.clarifai.com/blog/building-an-ai-app-with-clarifai-py… -
Clarifai
@clarifai
UI Modules are Clarifai’s integration with @streamlit to build AI web apps!
Here are 6 Modules for various Generative AI and Machine Learning applications:
- LLM Battleground
- Doc Q&A
- Chatbot Module
- Youtube LLM
- Data Augmentation
- Hyperparameter sweeps
Read more:🧵 ⬇️Post not available here, view on X
Clarifai
@clarifai
Data Augmentation:
This module augments data within a Clarifai application.
You can configure augmentation parameters such as horizontal flip, random pixel brightness, contrast. and others that you wish to apply per input image.
clarifai.com/clarifai/data/modules/data-augmentation
Clarifai
@clarifai
Hyperparameter sweeps:
This helps you experiment with a range of hyperparameter values and their combinations. You can define the range of values for a hyperparameter and set the step size for experimentation
clarifai.com/clarifai/data/modules/data-augmentation
Clarifai
@clarifai
That's a wrap!
Clarifai turns 10 on Nov 20th 🎉
This is Day 5 in this 10-day series and we keep sharing practical tips and resources to help you create production-grade AI.
Stay Tuned. -
Clarifai
@clarifai
The quality of input data directly impacts the model's performance.
Want to improve your efficiency while labeling your data? Here are 7 tips for efficient data labeling!
Check this out: ⬇️
www.clarifai.com/blog/7-tips-for-efficient-data-labeling -
Clarifai
@clarifai
8 tips to train and fine-tune machine learning models on Clarifai Platform!
Want to create your own model, fine-tune an existing one, or get started right away by using one of the pre-trained models?
Check this out: 👇
www.clarifai.com/blog/8-tips-to-train-and-fine-tune-machi… -
Clarifai
@clarifai
Workflows in Clarifai allows you to create robust multi-modal systems by combining models and performing different operations in a single call.
Here are 9 Workflows that helps you build innovative multi-modal applications: ⬇️Post not available here, view on X
Clarifai
@clarifai
Check out this tutorial to learn more about building workflows for your own multi-modal use cases.
www.clarifai.com/blog/workflows-in-clarifai-a-tutorial
Clarifai
@clarifai
That's a wrap!
Clarifai turns 10 on Nov 20th 🎉
This is Day 9 of the 10-day series as we continue to share practical tips and resources that helps you create production-grade AI.
Stay Tuned! -
Clarifai
@clarifai
Top 10 Open Source Large Language Models (LLMs) in 2023!
Thread🧵 ⬇️Post not available here, view on X
Clarifai
@clarifai
Read more about the Top 10 Open Source LLMs along with the links to try them out here:
www.clarifai.com/blog/top-10-open-source-large-language-m…
Clarifai
@clarifai
That's a wrap!
Clarifai turns 10 on Nov 20th 🎉
In this 10-day series we are going to share the practical tips and resources that helps you create production-grade AI.
Stay Tuned! -
Clarifai
@clarifai
Join us in welcoming Jonathan Padgett to lead Clarifai's Public Sector Sales and governmental partnerships. An exciting addition to our team at a time when demand for a full stack AI platform is only getting hotter! 🔥🔥🔥 www.clarifai.com/press-release/clarifai-hires-jonathan-pa…
-
Clarifai
@clarifai
Ready to take on the #genai wave?
The Ultimate Generative AI Glossary is filled with terms you absolutely should know to fully understand the Generative AI capabilities. Arm yourself with the Glossary and lead your company to success with Generative AI.
www.clarifai.com/ebook/ultimate-generative-ai-glossary?hs… -
Clarifai
@clarifai
Introducing Embed-v3 the latest text embedding model from @cohere!
clarifai.com/cohere/embed/models/cohere-embed-english-v3_0
Embed-v3 excels in semantic search and RAG systems with enhanced content quality assessment and efficiency.
Read more: ⬇️
Clarifai
@clarifai
Here are some key points on Embed v3:
- SOTA performance on trusted benchmarks like MTEB and BEIR
- Robustness to noisy datasets
- Compressed embeddings to save on storage costs
- Better retrievals for RAG systems
Clarifai
@clarifai
You can now access these models in the Clarifai Platform:
Try out the Cohere Embed English model:
clarifai.com/cohere/embed/models/cohere-embed-english-v3_0
Try out the Cohere Embed Multilingual model:
clarifai.com/cohere/embed/models/cohere-embed-multilingua… -
Clarifai
@clarifai
Want to understand how Large Language Models work?
The article helps you understand the core concepts in an interactive way!
One of the most beautiful visualizations on how LLMs work.
Check this out:
ig.ft.com/generative-ai/ -
Clarifai
@clarifai
Stanford CS224N: NLP with Deep Learning course from Stanford!
The course covers topics ranging from basics of NLP to advanced concepts like Attention and Transformer networks.
Inarguably, one of the best NLP courses on the web!
Check this out:
youtube.com/playlist?list=PLoROMvodv4rMFqRtEuo6SGjY4XbRIVRd4 -
Clarifai
@clarifai
Evaluating Large Language Models: A Comprehensive Survey on evaluating LLMs! 🔥
LLMs have attracted significant attention in recent past but they also suffer from potential risks like private data leaks or yield inappropriate, harmful, or misleading content.
This is a great survey that shows the different types of datasets and techniques to evaluate LLMs
Clarifai
@clarifai
Check this out:⬇️
arxiv.org/abs/2310.19736 -
Clarifai
@clarifai
Introducing SSD-1B: Fastest text-to-image model 🔥
clarifai.com/segmind/segmind-stable-diffusion/models/ssd-1b
Segmind Stable Diffusion 1B is the latest open-source text-to-image model which is 50% smaller and 60% faster than SDXL 1.0!
Python Code to try out SSD-1B: ⬇️ -
Clarifai
@clarifai
“A Survey of Large Language Models” is the most comprehensive survey paper on LLMs! 🔥
The paper provides an up-to-date review of the literature on LLMs, details about the training mechanisms like pre-training approaches along with instruction tuning techniques & further alignment training with the RLHF approach.
This also includes history and evolution of LLMs, useful prompting tips, evaluation of capabilities, a GitHub repo, and future directions.
Clarifai
@clarifai
Check this out: ⬇️
arxiv.org/abs/2303.18223 -
Clarifai
@clarifai
Thank you to everyone for being part of Clarifai's Generative AI Showdown in the AI DevWorld Hackathon!
There were so many great projects built 😍
Keep reading to find out the top 3 winners! 👇
1/4Post not available here, view on X
Clarifai
@clarifai
3. Facets Wellness App
This app helps you navigate hard life experiences by effectively analyzing your emotions and thoughts via journaling backed by psychology, coaching, neuroscience, etc.
devpost.com/software/hackjournal
Clarifai
@clarifai
Follow us @clarifai for more of these Hackathons around Generative AI!
Join the Community to stay up to date around LLMs!
discord.gg/ynDD6WKdb9 -
Clarifai
@clarifai
A Machine Learning course from Stanford taught by Andrew Ng!
You'll learn:
• Linear/Logistic regression
• Decision Trees
• KNN
• Support Vector Machines
• Neural Networks
• Naive Bayes
• Gradient Descent
Check this out:
youtube.com/playlist?list=PLoROMvodv4rMiGQp3WXShtMGgzqpfVfbU -
Clarifai reposted
Sarah Nagy
@sarahrnagy
if anyone is at ScaleUp AI NYC today please let me know! i had the great fortune to meet some absolute legends in our space such as @mattzeiler from @clarifai. thank you so much @insightpartners for throwing such a fabulous event!
#generativeai #dataanalytics -
Clarifai
@clarifai
Introducing GPT-Engineer an AI agent that can write an entire codebase with a single prompt!
It can also learns how you want your code to look.
Read more: ⬇️
Clarifai
@clarifai
More details of GPT-Engineer:
- Single prompt to codebase
- It asks you for clarifying questions
- Writes all the necessary code
- Easy to add your own reasoning steps, modify, and experiment
Github Repo: github.com/AntonOsika/gpt-engineer -
Clarifai
@clarifai
With a single line of text you can output the entire process of a software company along with carefully orchestrated SOPs 🔥
Introducing MetaGPT a Multi-Agent Framework that can act as an engineer, product manager, architect and project manager.
Clarifai
@clarifai
Github Repo: github.com/geekan/MetaGPT
-
Clarifai
@clarifai
LLMs Practical Guide!🔥
Here is the evolution Tree of Large Language Models from 2018 to 2023!
This Includes guide and resources to GPT-style Language Models:
→ GPT
→ GPT-2
→ GPT-3
→ GPT-4
→ OPT
→ PaLM
→ BLOOM
→ LaMDA
→ LLaMA 2
→ BloombergGPT
→ GPT-NeoX-20B
Also to the BERT-style Language Models:
→ BERT
→ RoBERTa
→ DistilBERT
→ ALBERT
→ UniLM
→ ELECTRA
→ T5
→ GLM
→ AlexaTM
→ ST-MoE
This also contains the Practical Guide of these models for the various NLP Tasks!
github.com/Mooler0410/LLMsPracticalGuide
Clarifai
@clarifai
You can access diverse Open Source and 3rd party LLMs from the Clarifai Community!
Check this out: 👇
clarifai.com/explore/models?filterData=[{"field"%3A"use_c… -
Clarifai
@clarifai
Google has released 7 free courses that helps you learn AI from scratch!
- Generative AI & LLMs fundamentals
- Diffusion Models
- Transformer and BERT Models
- Attention Mechanism
- Image Captioning Models
Thread:🧵 ⬇️
Clarifai
@clarifai
1. Introduction to Generative AI
cloudskillsboost.google/course_templates/536
2. Introduction to Large Language Models
cloudskillsboost.google/course_templates/539
3. Introduction to Responsible AI
cloudskillsboost.google/course_templates/554
4. Introduction to Image Generation:
www.cloudskillsboost.google/course_templates/541
Clarifai
@clarifai
5. Encoder-Decoder Architecture
cloudskillsboost.google/course_templates/543
6. Attention Mechanism
cloudskillsboost.google/course_templates/537
7. Transformer Models and BERT Model
www.cloudskillsboost.google/course_templates/538
Clarifai
@clarifai
That's a wrap! Follow us @clarifai for more content.
Join the Community below to talk about LLMs, Machine Learning and Generative AI!
discord.gg/yCzAdgZMkx -
Clarifai
@clarifai
Introducing SightCom an AI app designed for visually impaired individuals!
SightCom offers 5 unique features, all activated activated through voice commands:
1. Scene Description
2. Color Recognition
3. Text Recognition
4. Acts as a Product Reader &
5. Q&A Chatbot
Thread🧵⬇️
Clarifai
@clarifai
This is the winning Project in the Llama 2 Hackathon with Clarifai!
SightCom uses following models from Clarifai in the backend.
→ Blip 2,
→ Color Recognition model ,
→ OCR model &
→ Llama 2 7B.
Read more about the project in the blog below:
www.clarifai.com/blog/llama-2-hackathon-with-clarifai-win…
Clarifai
@clarifai
Checkout the Project here:
github.com/Louisljz/SightCom -
Clarifai
@clarifai
Join us to learn from the Winners of the @streamlit LLM Hackathon starting in 2 hours! 📣
Link: discord.gg/8JtvKECs?event=1163392197786349608
The 2 winners AI Research Buddy and AI Persona Simulator will be presenting their ideas and how they approached in building these projects! -
Clarifai
@clarifai
Introducing Microsoft's Autogen a framework that allows LLM agents to chat with each other to solve your tasks 🔥
AutoGen agents are customizable, conversable, and seamlessly allow human participation.
Read more: ⬇️
- AutoGen supports diverse conversation patterns for complex workflows and can be used to build a wide range of conversation patterns.
- AutoGen helps in building LLM Applications using multi-agent conversations.
- It simplifies the orchestration, automation, and optimization of a complex LLM workflow.
- It's also a drop-in replacement of openai.Completion or openai.ChatCompletion as an enhanced inference API.
Check this out: github.com/microsoft/autogen -
Clarifai
@clarifai
The recent LLM from @huggingface zephyr-7b-alpha outperforms Llama 2 70B Chat on MT Bench 🤯
clarifai.com/huggingface-research/zephyr/models/zephyr-7B…
You can now run zephyr-7b-alpha with just a few lines of code:
Read more: ⬇️
Clarifai
@clarifai
Running Zephyr 7B using Clarifai's Python SDK:
You can run the Zephyr 7B Alpha Model using Clarifai’s python SDK with just a few lines of code.
Check out the Code Below:
gist.github.com/Sumanth077/d50898800b2204be8427b09f4599b400
Clarifai
@clarifai
Checkout the model notes below to learn more about the Zephyr 7B Alpha model.
Also check out this to run the model with other API clients like Javascript, cURL, NodeJS, PHP, etc.
www.clarifai.com/blog/run-zephyr-7b-with-an-api -
Clarifai
@clarifai
ChatGPT vision but OpenSource 🔥
Introducing LLaVa, a 13B Multimodal LLM with vision assistance providing impressive chat capabilities similar to multimodal ChatGPT!
Clarifai
@clarifai
Learn more about it here: llava-vl.github.io
Follow us @clarifai for more content on LLMs and Generative AI! -
Clarifai
@clarifai
Join us for the Project Showcase of Clarifai Winners in the @streamlit LLM Hackathon!🔥
Link: ➡️ discord.gg/8JtvKECs?event=1163392197786349608
The 2 winners are:
1. AI Research Buddy &
2. AI Persona Simulator
Teams will be presenting their ideas and how they approached in building these projects. -
Clarifai
@clarifai
We are beyond excited about the AI Devworld Hackathon 2023! 🚀 🤖
Leverage Clarifai and Participate in the Generative AI Showdown
October 9 - 26, 2023
api-world-hackathon-2023.devpost.com
Join us to collaborate, innovate, and win exciting prizes:
Read more: 🧵 ⬇️Post not available here, view on X
Clarifai
@clarifai
We are offering these amazing Prizes: 🏆
1st Place: $500 Amazon gift card, Essential Plan Free for 1 year (worth $360), swag kit
2nd Place:$350 Amazon gift card, Essential Plan Free for 6 months (worth $180), swag kit
3rd Place: $150 Amazon gift card, swag kit
Clarifai
@clarifai
Resources: 🛠️
Join the aidevworld-generative-ai-showndown channel on our Discord for support.
The Clarifai team will be available to answer questions and guide you as you build your apps!
discord.com/channels/1145701543228735582/1161390096298283148 -
Clarifai
@clarifai
Introducing the Clarifai's New Python SDK! 🔥
You can now:
• Create apps, datasets, and do model predictions.
• Access models and LLMs from the Clarifai community with just a few lines of code.
Tutorial to help you get started with Python SDK: ⬇️
youtu.be/luF0jdptg-8 -
Clarifai
@clarifai
Introducing StreamingLLM 🔥
A framework that enable Large Language Models(LLMs) to handle infinite sequence lengths without the need for fine-tuning.
Thread🧵 ⬇️
Clarifai
@clarifai
Here is how StreamingLLM works:
One of the challenges of handling infinite length text is the Storage.
LLMs needs to stores all the Key and Value states that demands significant memory, and models might struggle to generate text beyond their training sequence length.
Clarifai
@clarifai
With StreamingLLM it addresses this by retaining only the most recent tokens and attentions, discarding the intermediate tokens.
This enables the model to generate coherent text from recent tokens without a cache reset.
Remember that the context window of LLM remains unchanged.
Clarifai
@clarifai
Read the Paper here: arxiv.org/abs/2309.17453
Github: github.com/mit-han-lab/streaming-llm -
Clarifai
@clarifai
Introducing the new Clarifai + @llama_index integration!
You can now access all the LLMs and embedding models from Clarifai Community through LlamaIndex!
docs.llamaindex.ai/en/latest/examples/llm/clarifai.html
Read more:🧵 ⬇️
Clarifai
@clarifai
Clarifai offers Open Source LLMs like Llama 2, Falcon, Stable Diffusion and the other 3rd party LLMs from Providers like Open AI, Cohere etc.
With this integration you can now access these from the Community!
Check out the Notebook below to get started:
docs.llamaindex.ai/en/latest/examples/llm/clarifai.html
Clarifai
@clarifai
Clarifai also offers Embedding Models that helps you create embeddings for your data.
You can now access these with just a few lines of code through LlamaIndex!
Check out the below example:
docs.llamaindex.ai/en/latest/examples/embeddings/clarifai… -
Clarifai
@clarifai
Here are the Top 2 Projects in the recent LLM Hackathon by @streamlit built with Clarifai! 🎉
1. AI Research Buddy which is a Q&A and summarization bot using the lastest Nougat transformer model for Academic papers.
github.com/bp-high/research_buddy
Clarifai
@clarifai
2. AI Persona Simulator which consists of 5 Different Personas that are capable producing art work, blog posts, and social media posts:
The Project uses Stable Diffusion XL & GPT 4 models from Clarifai Community.
Here are more details of the projects:
github.com/msull/consciousness-sim
Clarifai
@clarifai
Congratulations to all the Winners. Also thanks to @streamlit for organizing the Hackathon.
-
Clarifai
@clarifai
Introducing Nougat from @Meta!
A Visual Transformer model that converts document images including complex math equations, into a structured text!
clarifai.com/facebook/nougat/models/nougat-base
Thread ⬇️Post not available here, view on X
Clarifai
@clarifai
Model details: Decoder
The decoder then generates the output text tokens autoregressively using self-attention over the encoder outputs.
Nougat is trained end-to-end on page image and text pairs using stochastic gradient descent.
Clarifai
@clarifai
Nougat Model has a wide range of applications in the field of document understanding and extraction.
Some key use cases include Research Paper Parsing, Data Extraction and Summarization.
Try out the Model here: clarifai.com/facebook/nougat/models/nougat-base -
Clarifai reposted
LlamaIndex 🦙
@llama_index
LlamaIndex + @clarifai
We now extensive integrations with both @clarifai LLMs and embedding models.
LLMs: 40+ models from GPT-4, cohere, llama2, wizardcoder, etc.
Embeddings: access to BAAI-bge, multilingual-text-embedding, and more.
Huge thanks to @isaacchung1217! See 👇 -
Clarifai reposted
Streamlit
@streamlit
The two winning #llm hackathon apps using @clarifai:
🧍 Persona Simulator by Matthew Sullenberger: personas.streamlit.app/-/login?payload=MTc5NDQwMjYwMXxFST…
📖 AI Research Buddy by Bhavish Pahwa: gpt-research-buddy.streamlit.app/-/login?payload=MTc5NDQw…
3/7 -
Clarifai
@clarifai
Mistral 7B Instruct is available now!
Mistral 7B is the latest model from @MistralAI that outperforms Llama 2 13B on every benchmark.
Mistral 7B is finetuned on publicly available instruction datasets resulting in Mistral 7B Instruct
clarifai.com/mistralai/completion/models/mistral-7B-Instruct
Read more: ⬇️
Clarifai
@clarifai
You can run the model using the Python SDK with just a few lines of code:
Signup to Clarifai www.clarifai.com to get your personal access token (PAT) and run the below code
Check this out:
gist.github.com/Sumanth077/88a98a894a4be88e956139f1b0ac642c
Clarifai
@clarifai
As mentioned the model outperforms Llama 2 13B on every benchmark.
Key Features:
- It released under Apache 2.0 licence.
- Superior to Llama 34B on many benchmarks
- Approaches CodeLlama 7B performance on code
Checkout the Mistral 7B model notes here: clarifai.com/blog/run-mistral-7b-with-an-api -
Clarifai
@clarifai
Introducing WizardCoder a Large Language Model for Code that has been fine-tuned on Llama2!
WizardCoder also has it's finetuned version WizardCoder-Python-34B especially for Python.
You can now access these models with an API.
Read more🧵👇Post not available here, view on X
Clarifai
@clarifai
The 34B model is not just a coding assistant; it is also capable of generating snippets for:
1. Automating DevOps Scripts
2. Data Analysis
3. Generate End to End Machine Learning Pipelines
4. Web Scraping
6. Generate smart contracts for blockchain platforms
Clarifai
@clarifai
Try out the WizardCoder-15B model here: clarifai.com/wizardlm/generate/models/wizardCoder-15B
And the WizardCoder-Python-34B model:
clarifai.com/wizardlm/generate/models/wizardCoder-Python-34B
Clarifai
@clarifai
You can also run WizardCoder Models using an API with the Clarifai's Python client.
Read the blog here:
clarifai.com/blog/wizardcoder-large-language-model-for-code -
Clarifai
@clarifai
Open AI released GPT 3.5 Turbo Instruct Model that can be utilized for both conversation and for completing various tasks and questions.
Here is everything you need to know about the model:
Thread🧵 ⬇️Post not available here, view on X
Clarifai
@clarifai
Key Features and Advantages:
1. The primary strength of the model lies in its ability to follow instructions with precision and efficiency.
2. The model provides safer and more reliable AI interaction.
3. The model also demonstrated its prowess in the world of chess.
Clarifai
@clarifai
Try out the GPT-3.5 Turbo Instruct model here: 👇
clarifai.com/openai/completion/models/gpt-3_5-turbo-instruct -
Clarifai reposted
Clarifai
@clarifai
ElevenLabs leverages a powerful generative AI framework to convert written text into lifelike audio with remarkable accuracy.
You can now access the @elevenlabs model with just a few lines of python code.
Thread🧵 ⬇️ -
Clarifai
@clarifai
ElevenLabs leverages a powerful generative AI framework to convert written text into lifelike audio with remarkable accuracy.
You can now access the @elevenlabs model with just a few lines of python code.
Thread🧵 ⬇️Post not available here, view on X
Clarifai
@clarifai
• Language Learning: It also facilitates language learning by providing accurate pronunciation & intonation examples in various languages.
• Entertainment: Develop interactive and engaging games, interactive narratives, and virtual experiences with dynamic AI-generated voices.
Clarifai
@clarifai
Check the model in the Clarifai Platform.
Try it out: 👇
clarifai.com/eleven-labs/audio-generation/models/speech-s… -
Clarifai
@clarifai
Introducing the Clarifai's New Python SDK! 🔥
You can now:
• Import datasets along with annotations.
• Create apps, inputs, datasets, and do model predictions.
• Access models, workflows & other LLMs from the Clarifai community with just a few lines of code.
Read More:🧵👇Post not available here, view on X
Clarifai
@clarifai
Clarifai Workflows let's you combine multiple models and carry out different operations on the platform.
You can now even interact and run workflow predictions using the Python SDK.
Clarifai
@clarifai
Here is the Repository of the Python SDK.
Check this out and Star it to support the work: ⭐️
github.com/Clarifai/clarifai-python -
Clarifai
@clarifai
Introducing scikit-llm where Sklearn Meets Large Language Models.
scikit-llm allows you to integrate language models like ChatGPT into scikit-learn library to do text analysis.
Read more: ⬇️
github.com/iryna-kondr/scikit-llm
Clarifai
@clarifai
Here are all the tasks you can do with Scikit-LLM:
- Zero-Shot Text Classification
- Few-Shot Text Classification
- Dynamic Few-Shot Text Classification
- Text Classification with Google PaLM 2
- Text Vectorization
- LLM Fine-Tuning
- Text Summarization
- Text Translation
Clarifai
@clarifai
Learn more about it here:
github.com/iryna-kondr/scikit-llm -
Clarifai
@clarifai
If you are part of the @streamlit LLM Hackathon and building with Clarifai! 🚀
Join us today at 12 pm ET for the final Q&A session with the Clarifai team!
Link: discord.gg/EwESbQKm?event=1152332735042564097 -
Clarifai
@clarifai
With the increasing number of Large Language Models, it's really important to evaluate them accurately!
Here are the 5 most commonly used LLM evaluation frameworks and a short explanation to each:
Thread ⬇️Post not available here, view on X
Clarifai
@clarifai
4. MMLU: Massive Multitask Language Understanding
MMLU measures the knowledge acquired during the pretraining by evaluating models exclusively in zero-shot and few-shot settings.
It tests both the world knowledge of LLM and problem solving ability.
arxiv.org/abs/2009.03300
Clarifai
@clarifai
5. HellaSwag
HellaSwag tests the LLMs whether they can perform human level commonsense inference.
While humans find it easy (over 95% correct), the best SOTA models struggle (less than 48% correct).
Read more about it:
arxiv.org/abs/1905.07830 -
Clarifai
@clarifai
Don't miss out on office hours for the @streamlit LLM Hackathon on Discord!
This is a great opportunity to bring all the questions you have about building with Clarifai 🧑💻
When: Today, Sep. 14th at 12 pm EST
Join us here: discord.com/invite/VWYk4Kn9?event=1151152170842992681 -
Clarifai
@clarifai
Stable Diffusion XL 1.0 is the latest state-of-the-art latent diffusion model from @StabilityAI for high-resolution image synthesis.
You can now access and run Stable Diffusion XL 1.0 with an API.
Thread🧵⬇️Post not available here, view on X
Clarifai
@clarifai
SDXL can be used for various applications, including but not limited to:
• Text-to-image synthesis
• Image editing and manipulation
• Data augmentation for computer vision tasks
• Artistic image creation
Clarifai
@clarifai
Try out the Stable Diffusion XL 1.0 model here:
clarifai.com/stability-ai/stable-diffusion-2/models/stabl…
You can also run Stable Diffusion XL 1.0 Model using the Clarifai's Python client.
Read the blog here: clarifai.com/blog/run-stable-diffusion-xl-with-an-api -
Clarifai
@clarifai
LIVE now!
Our Product Marketing Manager @iankelk is hosting an "Ask the Clarifai Experts" for all the developers hacking with us for the @streamlit LLM Hackathon.
Join us now! discord.gg/edNJVzyv?event=1150895215293505656 -
Clarifai
@clarifai
StarCoder a Large Language Model for Code from @BigCodeProject!
This is trained on 80+ programming languages, Git commits, GitHub issues, and Jupyter notebooks.
Here are more details about StarCoder:🧵
Clarifai
@clarifai
StarCoder is a 15Billion parameter LLM for code with 8k context length which was trained on the licensed GitHub code.
Here are some other things the model can do:
- Writing code snippets
- Translate Python to C++
- Explain concepts
- Act as a terminal etc.
Clarifai
@clarifai
The model was trained on a subset of The Stack 1.2. This dataset only consists of licensed code.
Feel free to explore Star Coder in the Clarifai Platform!
Check this out:⬇️
clarifai.com/bigcode/code/models/StarCoder -
Clarifai
@clarifai
We're live right now!
Are you participating in the LLM Hackathon by @streamlit?
Join us for a live stream on “Working with Clarifai and LLMs”. 👀
discord.gg/kDNbgdkH?event=1149423204444688494 -
Clarifai
@clarifai
When dealing with large collections of digital assets, it's time consuming to sort through them all manually. Instead, using visual search, you can instantly find exactly what you're looking for. Our free e-book lays it all out!
#visualsearch www.clarifai.com/ebook/digital-asset-management -
Clarifai
@clarifai
WATCH LIVE! CEO @MattZeiler to talk #thefutureof AI on @BloombergLIVE w/ @alexbarinka @SentientDAI @UBSsmartwealth www.bloomberg.com/live/stream




















































































































































































































































































































































































































































































