3 projects
megatron-bridge
Megatron Bridge: Training Recipes for Megatron-based LLM and VLM models
nvFSDP
nvFSDP - NVIDIA fork of Fully Sharded Data Parallelism that is cross-compatible with Megatron / TransformerEngine and native PyTorch.
nemo-rl
NeMo RL: A Scalable and Efficient Post-Training Library for Models Ranging from 1 GPU to 1000s, and from Tiny to >100B Parameters
