Pinned Loading

1. TransformerEngine TransformerEngine Public

   Forked from NVIDIA/TransformerEngine

   A library for accelerating Transformer models on NVIDIA GPUs, including using 8-bit floating point (FP8) precision on Hopper and Ada GPUs, to provide better performance with lower memory utilizatio…

   Python 1

2. Megatron-LM Megatron-LM Public

   Forked from NVIDIA/Megatron-LM

   Ongoing research training transformer models at scale

   Python

3. Megatron-Bridge Megatron-Bridge Public

   Forked from NVIDIA-NeMo/Megatron-Bridge

   Training library for Megatron-based models with bidirectional Hugging Face conversion capability

   Python

4. RL RL Public

   Forked from NVIDIA-NeMo/RL

   Scalable toolkit for efficient model reinforcement

   Python