torchopt
torchopt copied to clipboard
metaopt
TorchOpt is an efficient library for differentiable optimization built upon PyTorch.
## Description Describe your changes in detail. ## Motivation and Context - [X] I have raised an issue to propose this change ([required](https://github.com/metaopt/TorchOpt/issues) for new features and bug fixes) Closes...
This is a proof of concept of integrating functorch in MAML.
## Description Add implicit MAML omniglot example ## Motivation and Context To test implicit differetiation API ## Checklist Go over all the following points, and put an `x` in all...
## Motivation Improve the speed of the current PyTree implemention: `tree_map`, `tree_flatten` and `unflatten` ## Solution An extra C++ library with python bindings ## Resource - https://github.com/pytorch/pytorch/issues/65761 - https://github.com/google/jax/issues/8099 ##...
Brand-new version of the old PR One can test this by running ```python python maml_torchrl.py --parallel --seed 1 ``` I have added the `--parallel` flag, to allow users to have...
## Motivation Task-level parallelization for multi-host multi-process optimization. Batch-level parallelization can be implemented easily by wrapping the network (`nn.Module`) with: - [`torch.nn.DataParallel`](https://pytorch.org/docs/stable/generated/torch.nn.DataParallel.html) (single-host multi-GPUs) (SPMD) - [`torch.nn.parallel.DistributedDataParallel`](https://pytorch.org/docs/stable/generated/torch.nn.parallel.DistributedDataParallel.html) (multi-host multi-GPUs) However,...
## Describe the bug A clear and concise description of what the bug is. ## To Reproduce Steps to reproduce the behavior. Please try to provide a minimal example to...
## Describe the bug CUDA part test failed on GPU machine ## Screenshots  ## System info  ```python import torchopt, numpy, sys print(torchopt.__version__, numpy.__version__, sys.version, sys.platform) 0.4.1 1.22.3 3.8.13...