🚀 PyTorch/XLA FSDP is supercharging the game in Hugging Face Transformers! Train PyTorch models with 20x more parameters using the same compute power🔥 Get ready for efficient GPT-2 training with up to 128B parameters on Google Cloud TPUs 👉 hubs.la/Q0207z760
2
17
74
23K
20
Download Image
