Reliable deep network training, saving you hours.
Project description
Reluble
A thin PyTorch wrapper which performs static and dynamic analysis of the computational graph to prevent latent errors.
Install
pip install reluble
Tests
Install the testing dependencies:
- pytest, e.g. with
pip install pytest
. - torchvision, e.g. with
pip install torchvision
. This is probably installed alongside torch.
Clone the repository and change into the root:
git clone git@github.com:arcadelab/reluble.git
cd reluble
To run all the tests, simply run pytest
. It is recommended to disable capturing, e.g. with:
pytest -s
These tests can take a long time, particularly if you are running on CPU. It is recommended to run
on GPU. Some of the tests will produce output files in ./outputs/
.
To show corruptions working for a good net, run:
pytest -s tests/test_training.py::TestCorruptions::test_good_Net
And for a bad net:
pytest -s tests/test_training.py::TestCorruptions::test_bad_Net
In this case, the LearningError
is the desired behavior.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
File details
Details for the file reluble-0.1.0.tar.gz
.
File metadata
- Download URL: reluble-0.1.0.tar.gz
- Upload date:
- Size: 12.5 kB
- Tags: Source
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.4.1 importlib_metadata/4.0.1 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.60.0 CPython/3.8.9
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | 5768597552ee77bf80f942bc54e1ea4938e7c42c30982c634b94d1743d1e6249 |
|
MD5 | a4561a705102798200e13720ffb42ee4 |
|
BLAKE2b-256 | ad14c8be5ba0d643b3e1c5e3688bf346d370dcc986ebd7931e4d27f7f417a976 |
File details
Details for the file reluble-0.1.0-py3-none-any.whl
.
File metadata
- Download URL: reluble-0.1.0-py3-none-any.whl
- Upload date:
- Size: 16.6 kB
- Tags: Python 3
- Uploaded using Trusted Publishing? No
- Uploaded via: twine/3.4.1 importlib_metadata/4.0.1 pkginfo/1.7.0 requests/2.25.1 requests-toolbelt/0.9.1 tqdm/4.60.0 CPython/3.8.9
File hashes
Algorithm | Hash digest | |
---|---|---|
SHA256 | f5077b69ecd92a62647dd41d9412d87e2eca99a732ac97c14a22a30b1fb778b8 |
|
MD5 | 9efc2bd0fe6c370cd51d66a29d61ee81 |
|
BLAKE2b-256 | 8ef23322dead12552dc6d2a949a93fbf6e3e65cb65be9c0badf709dba4234ba2 |