Sea AI Lab

Pinned Loading

1. understand-r1-zero understand-r1-zero Public

   Understanding R1-Zero-Like Training: A Critical Perspective

   Python 1.3k 63

2. zero-bubble-pipeline-parallelism zero-bubble-pipeline-parallelism Public

   Forked from NVIDIA/Megatron-LM

   Zero Bubble Pipeline Parallelism

   Python 466 32

3. lorahub lorahub Public

   [COLM 2024] LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition

   Python 670 44

4. oat oat Public

   🌾 OAT: A research-friendly framework for LLM online alignment, including reinforcement learning, preference learning, etc.

   Python 671 63

5. stde stde Public

   Official implementation of Stochastic Taylor Derivative Estimator (STDE) NeurIPS2024

   Python 128 9

6. feedback-conditional-policy feedback-conditional-policy Public

   Code for "Language Models Can Learn from Verbal Feedback Without Scalar Rewards"

   Python 65 2