Popular repositories Loading

1. ART ART Public

   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, and more!

   Python 6.6k 439

2. OpenPipe OpenPipe Public

   Turn expensive prompts into cheap fine-tuned models

   TypeScript 2.7k 156

3. deductive-reasoning deductive-reasoning Public

   Train your own SOTA deductive reasoning model

   Python 105 8

4. pii-redaction pii-redaction Public

   Detect and redact PII locally with SOTA performance

   Python 71 10

5. open_deep_research_training open_deep_research_training Public

   Training setup for Langchain's Open Deep Research

   Python 37 8

6. Summary-RL Summary-RL Public

   Train an agent to generate high quality summaries

   Jupyter Notebook 35 9