Last released Jul 29, 2024
Scalable GPT implementation using Distributed Data Parallel for efficient, multi-GPU training of transformer models.
Supported by