[CI] Add unit test workflows for Windows #804
Merged
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
unittest_windows_optdepts_cpu
workflow forpython_version=3.7, 3.8, 3.9, 3.10
andunittest_windows_optdepts_gpu
workflow forpython_version=3.7, 3.8, 3.9, 3.10
andcuda=116
(workflows are not run forcuda=117
as done in the Torchvision repo)Some unit tests are set to be skipped on Windows:
test/test_rb_distributed.py::test_funcs
:torch.distributed.rpc
support is unstable for Windowstest/test_rb.py:: TestPrototypeBuffers::{test_add,test_extend,test_sample,test_index}
withrb_type= RemoteTensorDictReplayBuffer
:torch.distributed.rpc
support is unstable for Windowstest_timeit.py
is taking more time than allowed on Windows:test_shared_.py::test_memmap
: Call totd.to(SavedTensorDict)
is failing (for all combinations ofdtype
andidx
) with traceback:test_collector.py::test_output_device_consistency
:Windows Access Violations
andBrokenPipeError
for certain combinations of python/policy_device/device/num_envtest_collector.py::test_collector_batch_size
: Test timeout (> 10 min) on CI pipeline Windows machine with GPU fornum_env=3
test_collector.py::test_collector_device_combinations
:Access violation
intorch.storage
Motivation and Context
TorchRL is currently built and tested for OsX and Linux only and there is some demand for a windows built.
This requires adding some test workflows for Windows.
Checklist