49c3587591
* chore(components/pytorch):kserve migration Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * fix: pytorch dist training - enable env vars in config.properties - upgrade pip in dockerfile Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * Bert - KServe v2 handler changes Signed-off-by: Shrinath Suresh <shrinath@ideas2it.com> * fix: bert notebook for kserve v2 Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * fix: add protocol verion to bert gpu yaml Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * Adding utility to convert image to bytes - Cifar Signed-off-by: Shrinath Suresh <shrinath@ideas2it.com> * Cifar10 - captum update Signed-off-by: Shrinath Suresh <shrinath@ideas2it.com> * fix: cifar10 example Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * fix: predictor component for kserve v2 Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * fix: pytorch dist training for kserve v2 Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * fix: cifar10 hpo example Signed-off-by: Jagadeesh J <jagadeeshj@ideas2it.com> * Bumping pytorch-kfp-components version Signed-off-by: Shrinath Suresh <shrinath@ideas2it.com> Co-authored-by: Shrinath Suresh <shrinath@ideas2it.com> |
||
|---|---|---|
| .. | ||
| pytorch_kfp_components | ||
| templates | ||
| tests | ||
| .gitignore | ||
| .pylintrc | ||
| LICENSE | ||
| OWNERS | ||
| README.md | ||
| pyproject.toml | ||
| setup.cfg | ||
| setup.py | ||
| tox.ini | ||
README.md
PyTorch Kubeflow Pipeline Components
PyTorch Kubeflow Pipeline Components provides an SDK and a set of components that lets you build kubeflow pipelines using PyTorch. You can use the predefined components in this repository to build your pipeline using the Kubeflow Pipelines SDK.
Installation
Requirements
Python >= 3.6 Kubeflow cluster setuo (on-prem or in any of the Clouds)
Install latest release
Use the following command to install PyTorch Pipeline Components from PyPI.
pip install -U pytorch-kfp-components
Install from source
Use the following commands to install PyTorch Kubeflow Pipeline Components from GitHub.
git clone https://github.com/kubeflow/pipelines.git
pip install pipelines/components/PyTorch/pytorch_kfp_components/.
Running the tests
Run the following command
pip install tox
cd ./components/PyTorch/pytorch-kfp-components/
tox -rvve py38
Samples
For running the samples follow the instruction mentioned as below