Starred repositories
Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen2.5, Qwen3, Llama, Kimi, and more!
SkyPilot: Run AI and batch jobs on any infra (Kubernetes or 16+ clouds). Get unified execution, cost savings, and high GPU availability via a simple interface.
Simple UI for debugging correlations of text embeddings
Interactive visualizations of the geometric intuition behind diffusion models.
A collection of MCP (Model Context Protocol) tools and examples for wandb and weave
The official Python SDK for Model Context Protocol servers and clients
Genome modeling and design across all domains of life
Interactive Seminars with Real-time Feedback and Gemini based Q&A
Synthetic data generation and benchmark implementation for "Episodic Memories Generation and Evaluation Benchmark for Large Language Models" [ICLR 2025]
This will contain all my learning notes. It can be from a book, paper, documentation or any learning resources.
2
Updated Feb 5, 2025
Notebooks for fine tuning pali gemma
Fully open reproduction of DeepSeek-R1
The official evaluation suite and dynamic data release for MixEval.
Bringing BERT into modernity via both architecture changes and scaling
A generative world for general-purpose robotics & embodied AI learning.
Papers and resources related to the security and privacy of LLMs 🤖
Supercharge Your LLM Application Evaluations 🚀
Agent Framework / shim to use Pydantic with LLMs
A full Python Implementation of the ROUGE Metric (not a wrapper)
Weave is a toolkit for developing AI-powered applications, built by Weights & Biases.
A collection of LogitsProcessors to customize and enhance LLM behavior for specific tasks.
PyTorch native quantization and sparsity for training and inference
Reinforcement learning on general 2D physics environments in JAX. ICLR 2025 Oral.
🤗 LeRobot: Making AI for Robotics more accessible with end-to-end learning
The entmax mapping and its loss, a family of sparse softmax alternatives.
DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models
Educational framework exploring ergonomic, lightweight multi-agent orchestration. Managed by OpenAI Solution team.
JAX Implementation of Black Forest Labs' Flux.1 family of models