Leverage PyTorch 2.0's torch.compile to enhance the training speed of Bayesian Flow Networks. The new function promises faster model execution by compiling models to take advantage of modern hardware accelerators. Implement torch.compile with the correct settings to avoid common pitfalls and achieve optimal training performance.
To leverage the use of torch.compile in SuperGradients, you need to pass the torch_compile: True option to training hyperparameters. - PyTorch 2.0 Overview and Guide
Speed Up Model Training using torch 2.0 features
Leverage PyTorch 2.0's
torch.compile
to enhance the training speed of Bayesian Flow Networks. The new function promises faster model execution by compiling models to take advantage of modern hardware accelerators. Implementtorch.compile
with the correct settings to avoid common pitfalls and achieve optimal training performance.tonystark-bot/Bayesian-Flow-Networks/train.py
🧊 Repo link: tonystark-bot/Bayesian-Flow-Networks 🤖 Comment generated by ccc.inc ❓ Reply with
/regenerate <feedback>
to improve the card, or/help
to see all commands.