Skip to main content
Yanked

This release has been yanked by its maintainers, and will be ignored by installers, except when explicitly specified.
Consider using release 1.19.2 instead.

training pytorch models with onnxruntime

torch-ort can be used to train pytorch models with onnxruntime backend. Before using torch-ort, one need to have a working pytorch gpu environment.

to build (you need to update version number in version.txt in order to be able to upload a python whl):

(run 'pip install setuptools', if it is not already installed) 
rm dist/*
python setup.py bdist_wheel

to publish (it will ask for user name and password):

twine upload dist/*

to install:

stable:

TBD

nightly:

(make sure you are not in the ort repo folder - otherwise torch-ort is taken as already installed)
pip install --pre --extra-index-url https://aiinfra.pkgs.visualstudio.com/PublicPackages/_packaging/ORT-Nightly/pypi/simple/ ort-gpu-nightly-training
pip install torch-ort

(eventually we are aiming at: pip install --pre torch-ort onnxruntime=1.9.0+cu111_training)

to test:

python ./ort/tests/bert_for_sequence_classification.py

to use torch-ort within PyTorch training scripts:

import onnxruntime
from torch_ort import ORTModule
model = ORTModule(model)
# normal PyTorch training script follows

FAQs

Question: When running training script with ORTModule, I got an error of missing Cuda shared library:

Answer: It is possible that your torch install has a different version of Cuda library than onnxruntime. In this case, do:

conda install -c anaconda cudatoolkit=10.2.89
(replace with the right cudatoolkit version)

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distributions

No source distribution files available for this release.See tutorial on generating distribution archives.

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

torch_ort-0.0.9.dev20210322-py3-none-any.whl (3.1 kB view details)

Uploaded Python 3

File details

Details for the file torch_ort-0.0.9.dev20210322-py3-none-any.whl.

File metadata

  • Download URL: torch_ort-0.0.9.dev20210322-py3-none-any.whl
  • Upload date:
  • Size: 3.1 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/3.4.1 importlib_metadata/3.7.3 pkginfo/1.7.0 requests/2.22.0 requests-toolbelt/0.9.1 tqdm/4.59.0 CPython/3.8.5

File hashes

Hashes for torch_ort-0.0.9.dev20210322-py3-none-any.whl
Algorithm Hash digest
SHA256 f4f408470f506408543dece333c6e24d1a017d378455da45a0218df8fd2f2955
MD5 6709221cea7a5825166fefbedbcc35df
BLAKE2b-256 fe01264f84f16b79dc1288bca021868c605dc2fa14069a255cb69f508eee9ba5

See more details on using hashes here.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page