4970927ec8
* auto batch finder * fix styling * add description * add different modes * fix copy paste error * better organised code * fix styling * add tests * fix * fix * add some documentation * added CHANGELOG.md * some documentation * update based on review * Update trainer.py * Update docs/source/training_tricks.rst Co-authored-by: Adrian Wälchli <aedu.waelchli@gmail.com> * Update tests/trainer/test_trainer_tricks.py Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com> * Update tests/trainer/test_trainer_tricks.py Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com> * Apply suggestions from code review Co-authored-by: Adrian Wälchli <aedu.waelchli@gmail.com> * use EvalModelTemplate * param tests * rename * wrap params * rename function * rename * rename param * fix * abs * rename * refactor code * add docs * try * arg * loop * exept * loop * drop bool * docs * docs * added check and test for passing dataloader to fit * styling fix * update based on review Co-authored-by: Nicki Skafte <nugginea@gmail.com> Co-authored-by: William Falcon <waf2107@columbia.edu> Co-authored-by: Adrian Wälchli <aedu.waelchli@gmail.com> Co-authored-by: Jirka Borovec <Borda@users.noreply.github.com> Co-authored-by: Jirka <jirka.borovec@seznam.cz>
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
The automatic travis tests ONLY run CPU-based tests. Although these cover most of the use cases, run on a 2-GPU machine to validate the full test-suite.
To run all tests do the following:
git clone https://github.com/PyTorchLightning/pytorch-lightning
cd pytorch-lightning
# install AMP support
bash tests/install_AMP.sh
# install dev deps
pip install -r tests/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.
- NVIDIA-apex installed.
- Horovod with NCCL support:
HOROVOD_GPU_ALLREDUCE=NCCL HOROVOD_GPU_BROADCAST=NCCL pip install horovod
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 tests/requirements-devel.txt)
coverage run --source pytorch_lightning -m py.test pytorch_lightning tests examples -v --doctest-modules
# 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-pt_1_4 -f tests/Dockerfile --build-arg TORCH_VERSION=1.4 .
To build other versions, select different Dockerfile.
docker image list
docker run --rm -it pytorch_lightning:devel-pt_1_4 bash
docker image rm pytorch_lightning:devel-pt_1_4