ML Frameworks: Hugging Face Accelerate w/ Sylvain Gugger

In this edition of ML Frameworks W&B’s host Aman Arora was joined by HF Research Engineer, Sylvain Gugger. --- HF Accelerate was created for PyTorch users who like to write the training loop of PyTorch models but are reluctant to write and maintain the boilerplate code needed to use multi-GPUs/TPU/fp16. Accelerate abstracts exactly and only the boilerplate code related to multi-GPUs/TPU/fp16 and leaves the rest of your code unchanged. 0:00 Welcome & Intro 03:13 Sylvain Gugger on Hugging Face Study Groups on Deep Learning & Transformer Libraries 06:30 Sylvain Gugger on Hugging Face Accelerate 08:26 Distributed Training in Pytorch 41:15 Live Training Example on HF Accelerate 45:51 HF Accelerate Library 57:08 Q&A on HF Accelerate