saymrwulf/onnxruntime
1
0
Fork 0
mirror of https://github.com/saymrwulf/onnxruntime.git synced 2026-05-14 20:48:00 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Update docs/Model_Test.md (#11466)

Browse source 
Update the instructions of how to get test models.
This commit is contained in:
Changming Sun 2024-05-15 11:33:11 -07:00 committed by GitHub
parent 2a17958b34
commit ee3f2f4ebf
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1 changed files with 13 additions and 20 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

33
docs/Model_Test.md
View file

@ -1,27 +1,20 @@
ONNX has a collection of standard tests. This document describes how to run these tests through a C++ program named 'onnx_test_runner' in this repo. You could also run these test through onnxruntime python binding, which would be much easier to setup, but, a bit harder to debug issues. ONNX has a collection of standard tests. This document describes how to run these tests through a C++ program named 'onnx_test_runner' in this repo. You could also run these test through onnxruntime python binding, which would be much easier to setup, but, a bit harder to debug issues.
# Get the test data # Get the test data
You should have: ```
1. onnx single node test data git submodule update --init --recursive
2. onnx model zoo models pushd .
cd cmake/external/emsdk
./emsdk install latest
./emsdk activate latest
source ./emsdk_env.sh
popd
cd js
npm install
npm run prepare-node-tests
```
## Install onnx python package In addition to that, You can get more test models with their test data from https://github.com/onnx/models .
You can get onnx python package from [pypi](https://pypi.org/). However, if you are a onnxruntime developer, you may need to work on a cutting edge ONNX version. In this case, you need to build and install ONNX from source code.
### Install ONNX from source code
1. (windows) set ONNX_ML=1
(linux) export ONNX_ML=1
2. Install protobuf and put protoc into your PATH environment. When you compile protobuf, it's better to only enable the static libraries.
3. run "python setup.py bdist_wheel" and "pip install dist/*.whl"
## Generate node test data
$ python3 -m onnx.backend.test.cmd_tools generate-data -o <dest_folder>
e.g.
python3 -m onnx.backend.test.cmd_tools generate-data -o C:\testdata
## Get more models
Download https://onnxruntimetestdata.blob.core.windows.net/models/20190419.zip and unzip it.
# Compile onnx_test_runner and run the tests # Compile onnx_test_runner and run the tests