Public functional modules for Keras, TF and PyTorch
CircleCI is used for CI system:
This library consists of a few pretty much independent submodules:
| Module | Name | Info |
|---|---|---|
| Elastic | elastic |
Make Keras and PyTorch models elastic |
| Gradient Accumulation | ga |
Gradient accumulation for Keras and PyTorch optimizers |
| HPO | hpo |
Hyperparameter optimization assistance |
| Model Parallelism | mp |
Model-parallelism support for Keras builtin layers |
| Auto Profiler | profiler |
Export timeline of TF/Keras models easily |
| Reporter | reporter |
An interface to send metrics and parameters to Promethues Push Gateway |
Install the runai Python library using pip using the following command:
pip install runai
Make sure to use the correct
pipinstaller (you might need to usepip3for Python3)
All tests (unit tests) can be run using the following command:
python -m unittest discover -s tests -v