search

TorchJD / torchjd

Library for Jacobian descent with PyTorch. It enables optimization of neural networks with multiple losses (e.g. multi-task learning).
https://torchjd.org
MIT License
130 stars 0 forks source link

Improve documentation of parallel_chunk_size #107

Closed ValerianRey closed 2 weeks ago

ValerianRey commented 2 weeks ago

Very small improvement to the documentation that states explicitly the current limitation of the retain_graph parameter (corresponding to case D in the table of this issue).

You can merge if you agree.