An open API service indexing awesome lists of open source software.

https://github.com/sdpython/experimental-experiment

investigations
https://github.com/sdpython/experimental-experiment

Last synced: 5 months ago
JSON representation

investigations

Awesome Lists containing this project

README

        

.. image:: https://github.com/sdpython/experimental-experiment/raw/main/_doc/_static/logo.png
:width: 120

experimental-experiment: fuzzy work
===================================

.. image:: https://dev.azure.com/xavierdupre3/experimental-experiment/_apis/build/status/sdpython.experimental-experiment
:target: https://dev.azure.com/xavierdupre3/experimental-experiment/

.. image:: https://badge.fury.io/py/experimental-experiment.svg
:target: http://badge.fury.io/py/experimental-experiment

.. image:: http://img.shields.io/github/issues/sdpython/experimental-experiment.png
:alt: GitHub Issues
:target: https://github.com/sdpython/experimental-experiment/issues

.. image:: https://img.shields.io/badge/license-MIT-blue.svg
:alt: MIT License
:target: https://opensource.org/license/MIT/

.. image:: https://img.shields.io/github/repo-size/sdpython/experimental-experiment
:target: https://github.com/sdpython/experimental-experiment/
:alt: size

.. image:: https://img.shields.io/badge/code%20style-black-000000.svg
:target: https://github.com/psf/black

.. image:: https://codecov.io/gh/sdpython/experimental-experiment/branch/main/graph/badge.svg?token=Wb9ZGDta8J
:target: https://codecov.io/gh/sdpython/experimental-experiment

Getting started
+++++++++++++++

pytorch nightly build should be installed, see
`Start Locally `_.

::

git clone https://github.com/sdpython/experimental-experiment.git
pip install onnxruntime-gpu pynvml
pip install -r requirements-dev.txt
export PYTHONPATH=$PYTHONPATH:

Compare torch exporters
+++++++++++++++++++++++

The script evaluates the memory peak, the computation time of the exporters.
It also compares the exported models when run through onnxruntime.
The full script takes around 20 minutes to complete. It stores on disk
all the graphs, the data used to draw them, and the models.

::

python _doc/examples/plot_torch_export.py -s large

Documentation
+++++++++++++

See `experimental-experiment `_.