Pinned Loading

1. CUDA_Flash_Attention2 CUDA_Flash_Attention2 Public

   Implement Flash Attention v2 just from the paper in Numba JIT and CUDA

   Cuda 2

2. SmolLm2-zero SmolLm2-zero Public

   Forked from philschmid/deep-learning-pytorch-huggingface

   Train a small LLM to "think" with not SFT, only RL

   Jupyter Notebook 2

3. CleanGPT CleanGPT Public

   simple GPT2 pytorch implementation and pre-training on edu-Fineweb dataset

   Python