distributed-llm-pretraining-torchtitan

【效率与协作】Provides PyTorch-native distributed LLM pretraining using torchtitan with 4D parallelism (FSDP2, TP, PP, CP). Use when pretraining Llama 3.1, DeepSeek V3, or custom models at scale from 8 to 512+ GPUs with Float8, torch.compile, and distributed checkpointing.

独立 GitHub 快照 · 自动分类 · 以原作者文档为准