History

Akash Kwatra eff67d7a02 Deprecate `AbstractProfiler` in favor of `BaseProfiler` (#12106 )		2022-03-05 02:35:57 +00:00
..
accelerators	CI: update poplar sdk version (#12226 )	2022-03-04 23:49:30 +00:00
benchmarks	add parity test for sync batchnorm (#12021 )	2022-02-26 03:51:57 +00:00
callbacks	add `state_dict`/`load_state_dict` to base `Callback` (#11998 )	2022-03-04 02:41:48 +00:00
checkpointing	Stop loading a few properties if checkpoint's `dirpath` has changed (#12045 )	2022-02-28 16:42:09 +00:00
core	Fix `LightningModule.{un,}toggle_model` when only 1 optimizer is used (#12088 )	2022-02-28 12:41:51 +00:00
deprecated_api	Deprecate `AbstractProfiler` in favor of `BaseProfiler` (#12106 )	2022-03-05 02:35:57 +00:00
helpers	Refactor Horovod NCCL check (#11948 )	2022-02-28 10:45:32 +00:00
lite	Return the output of the optimizer step (#11711 )	2022-02-09 09:37:13 +00:00
loggers	Mark `logger_connector` as protected (#12195 )	2022-03-05 02:33:42 +00:00
loops	Improve mechanism to reset the seed after sanity check (#11870 )	2022-03-01 23:27:30 +00:00
models	Deprecate `weights_save_path` from the Trainer constructor (#12084 )	2022-02-28 22:45:26 +00:00
overrides	Fix retrieval of batch indices when dataloader num_workers > 0 (#10870 )	2021-12-02 10:36:10 +00:00
plugins	add `accelerator.is_available()` check (#12104 )	2022-03-02 10:07:49 +00:00
profiler	Deprecate `LoggerCollection` in favor of `trainer.loggers` (#12147 )	2022-03-04 23:01:43 +00:00
strategies	Check `parallel_devices` passed through `strategy` is consistent with the `accelerator` flag (#12105 )	2022-03-03 10:30:24 -08:00
trainer	Remove `data_pipeline` attribute patch (#12204 )	2022-03-04 23:09:37 +00:00
tuner	Update `tests/tuner/*.py` to use `devices` instead of `gpus` or `ipus` (#11520 )	2022-02-03 20:58:13 +05:30
utilities	Improve mechanism to reset the seed after sanity check (#11870 )	2022-03-01 23:27:30 +00:00
README.md	CI: add mdformat (#8673 )	2021-08-03 18:19:09 +00:00
__init__.py	Replace `yapf` with `black` (#7783 )	2021-07-26 13:37:35 +02:00
conftest.py	Add support for pluggable Accelerators (#12030 )	2022-02-28 21:36:23 +05:30
standalone_tests.sh	Fix selection of standalone tests (#10857 )	2021-12-01 09:48:37 +01:00

README.md

PyTorch-Lightning Tests

Most PL tests train a full MNIST model under various trainer conditions (ddp, ddp2+amp, etc...). This provides testing for most combinations of important settings. The tests expect the model to perform to a reasonable degree of testing accuracy to pass.

Running tests

git clone https://github.com/PyTorchLightning/pytorch-lightning
cd pytorch-lightning

# install dev deps
pip install -r requirements/devel.txt

# run tests
py.test -v

To test models that require GPU make sure to run the above command on a GPU machine. The GPU machine must have at least 2 GPUs to run distributed tests.

Note that this setup will not run tests that require specific packages installed such as Horovod, FairScale, NVIDIA/apex, NVIDIA/DALI, etc. You can rely on our CI to make sure all these tests pass.

Running Coverage

Make sure to run coverage on a GPU machine with at least 2 GPUs and NVIDIA apex installed.

cd pytorch-lightning

# generate coverage (coverage is also installed as part of dev dependencies under requirements/devel.txt)
coverage run --source pytorch_lightning -m py.test pytorch_lightning tests examples -v

# print coverage stats
coverage report -m

# exporting results
coverage xml

Building test image

You can build it on your own, note it takes lots of time, be prepared.

git clone <git-repository>
docker image build -t pytorch_lightning:devel-torch1.9 -f dockers/cuda-extras/Dockerfile --build-arg TORCH_VERSION=1.9 .

To build other versions, select different Dockerfile.

docker image list
docker run --rm -it pytorch_lightning:devel-torch1.9 bash
docker image rm pytorch_lightning:devel-torch1.9