Blog, Articles, and discussions

NEW 你也可以阅读这篇博客的中文版

Community Articles

ViDoRe V3: a comprehensive evaluation of retrieval for enterprise use-cases

and 4 others •

The 1 Billion Token Challenge: Finding the Perfect Pre-training Mix

Granite 4.0 Nano: Just how small can you go?

and 1 other •

Why Did MiniMax M2 End Up as a Full Attention Model?

⚡ Power, Heat, and Intelligence ☁️ - AI Data Centers Explained 🏭

and 1 other •

Llasa Goes RL: Training LLaSA with GRPO for Improved Prosody and Expressiveness

Evaluate Your Own RAG: Why Best Practices Failed Us

and 4 others •

What makes good reasoning data

KV Caching Explained: Optimizing Transformer Inference Efficiency

Ultra-Long Sequence Parallelism: Ulysses + Ring-Attention Technical Principles and Implementation

Budget Alignment: Making Models Reason in the User’s Language

and 2 others •

🌳 QAT: The Art of Growing a Bonsai Model

Effective Prompting for Generative Vision Models

and 1 other •

Introduction to State Space Models (SSM)

Small Language Models (SLM): A Comprehensive Overview

Post-Training Isaac GR00T N1.5 for LeRobot SO-101 Arm

and 5 others •

From GRPO to DAPO and GSPO: What, Why, and How

On the Shifting Global Compute Landscape

and 1 other •

Who Routes LLM Routers? RouterArena: Building the Evaluation Foundation for LLM Routing

and 6 others •

about 3 hours ago

SYNTH: the new data frontier

No GPU left behind: Unlocking Efficiency with Co-located vLLM in TRL

+2

agentssmolagentsopen-source

CodeAgents + Structure: A Better Way to Execute Actions

🐯 Liger GRPO meets TRL

+2

dellenterprisepartnerships

Dell Enterprise Hub is all you need to build AI on premises

+4

agentshubinference

Tiny Agents in Python: a MCP-powered agent in ~70 lines of code

guidediffusersquantization

Exploring Quantization Backends in Diffusers

nanoVLM: The simplest repository to train your VLM in pure PyTorch

+3

partnershipannouncementazure

Microsoft and Hugging Face expand collaboration

communityannouncementtransformers

The Transformers Library: standardizing model definitions

communityannouncement

Improving Hugging Face Model Access for Kaggle Users

+1

guidecloudinference

Blazingly fast whisper transcriptions with Inference Endpoints

+2

vlmvisionmultimodal

Vision Language Models (Better, faster, stronger)

+1

lerobotcommunitydatasets

LeRobot Community Datasets: The “ImageNet” of Robotics — When and How?

+3

How to Build an MCP Server with Gradio

Community Articles

ViDoRe V3: a comprehensive evaluation of retrieval for enterprise use-cases

and 4 others •

The 1 Billion Token Challenge: Finding the Perfect Pre-training Mix

Granite 4.0 Nano: Just how small can you go?

and 1 other •

Why Did MiniMax M2 End Up as a Full Attention Model?

⚡ Power, Heat, and Intelligence ☁️ - AI Data Centers Explained 🏭

and 1 other •

Llasa Goes RL: Training LLaSA with GRPO for Improved Prosody and Expressiveness

Evaluate Your Own RAG: Why Best Practices Failed Us

and 4 others •

What makes good reasoning data

KV Caching Explained: Optimizing Transformer Inference Efficiency

Ultra-Long Sequence Parallelism: Ulysses + Ring-Attention Technical Principles and Implementation

Budget Alignment: Making Models Reason in the User’s Language

and 2 others •

🌳 QAT: The Art of Growing a Bonsai Model

Effective Prompting for Generative Vision Models

and 1 other •

Introduction to State Space Models (SSM)

Small Language Models (SLM): A Comprehensive Overview

Post-Training Isaac GR00T N1.5 for LeRobot SO-101 Arm

and 5 others •

From GRPO to DAPO and GSPO: What, Why, and How

On the Shifting Global Compute Landscape

and 1 other •

Who Routes LLM Routers? RouterArena: Building the Evaluation Foundation for LLM Routing

and 6 others •

about 3 hours ago

SYNTH: the new data frontier

View all