torchopt
torchopt copied to clipboard
TorchOpt is an efficient library for differentiable optimization built upon PyTorch.
## Description make `root_vjp` public to users ## Types of changes What types of changes does your code introduce? Put an `x` in all the boxes that apply: - [x]...
### Required prerequisites - [X] I have searched the [Issue Tracker](https://github.com/metaopt/torchopt/issues) and [Discussions](https://github.com/metaopt/torchopt/discussions) that this hasn't already been reported. (+1 or comment there if it has.) - [X] Consider asking...
### Required prerequisites - [X] I have searched the [Issue Tracker](https://github.com/metaopt/torchopt/issues) and [Discussions](https://github.com/metaopt/torchopt/discussions) that this hasn't already been reported. (+1 or comment there if it has.) - [X] Consider asking...
### Required prerequisites - [X] I have read the documentation . - [X] I have searched the [Issue Tracker](https://github.com/metaopt/torchopt/issues) and [Discussions](https://github.com/metaopt/torchopt/discussions) that this hasn't already been reported. (+1 or comment...
## Description Describe your changes in detail. Explicitly wrap `torch.autograd.grad()` with `torch.enable_grad()` context. Users can use inference-only mode for hyper gradient with: ```python meta_optim = torchopt.MetaAdam(model, lr=0.1) loss = compute_loss(model,...
### Required prerequisites - [X] I have read the documentation . - [X] I have searched the [Issue Tracker](https://github.com/metaopt/torchopt/issues) and [Discussions](https://github.com/metaopt/torchopt/discussions) that this hasn't already been reported. (+1 or comment...
## Description when use torchopt.MetaAdam and step some times, the memory use in gpu are continuously increase. It should not be, will you excute next step, the tensor create in...
### Required prerequisites - [X] I have read the documentation . - [X] I have searched the [Issue Tracker](https://github.com/metaopt/torchopt/issues) and [Discussions](https://github.com/metaopt/torchopt/discussions) that this hasn't already been reported. (+1 or comment...
updates: - [github.com/pre-commit/pre-commit-hooks: v4.6.0 → v6.0.0](https://github.com/pre-commit/pre-commit-hooks/compare/v4.6.0...v6.0.0) - [github.com/pre-commit/mirrors-clang-format: v18.1.8 → v21.1.6](https://github.com/pre-commit/mirrors-clang-format/compare/v18.1.8...v21.1.6) - [github.com/astral-sh/ruff-pre-commit: v0.5.0 → v0.14.7](https://github.com/astral-sh/ruff-pre-commit/compare/v0.5.0...v0.14.7) - [github.com/PyCQA/isort: 5.13.2 → 7.0.0](https://github.com/PyCQA/isort/compare/5.13.2...7.0.0) - https://github.com/psf/black → https://github.com/psf/black-pre-commit-mirror - [github.com/psf/black-pre-commit-mirror: 24.4.2 → 25.11.0](https://github.com/psf/black-pre-commit-mirror/compare/24.4.2...25.11.0)...
Bumps [pypa/cibuildwheel](https://github.com/pypa/cibuildwheel) from 2.19 to 2.21. Release notes Sourced from pypa/cibuildwheel's releases. v2.21.0 ⚠️ Update CPython 3.12 to 3.12.6, which changes the macOS minimum deployment target on CPython 3.12 from...