-
Notifications
You must be signed in to change notification settings - Fork 360
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: Python Runtime Windows Builds on TRT 10 #2764
Conversation
120413b
to
8db8a0a
Compare
https://github.com/HolyWu/TensorRT/tree/windows_CI_trt_10_fix |
96f129e
to
6204ef8
Compare
Thanks very much @HolyWu - I've cherry-picked in the non-C++ changes (the C++ runtime will be ported in a different PR), plus some test modifications, and will see how the CI tests go. |
6204ef8
to
bc03fbf
Compare
bc03fbf
to
5fac363
Compare
Add build-test-windows job 9.3.0 Upgrade and version fixing Add bazelisk installation Update installation of Torch-TRT Bazel install fix Python only for initial Windows build Conda install latest cuda toolkit CUDA Driver Update CUDA Driver Update 2 Get windows version info Attempt driver upgrade within system Switch installer script to curl Switch installer script to network Try testing E2E models on Windows Try linux testing action for Windows Update Windows build/test Use local version of yml Use local version of yml 2 Use local version of yml 3 Use local version of yml 4 Remove CONDA_RUN references Update Python source Add torch-trt install script packing Remove extraneous python references Replace pipefail command Replace pipefail command 2 Replace source command Update cat command for Powershell Avoid script overwrite in runner Avoid script overwrite in runner 2 Update path concat to use echo fix: Use forked infra repository Skip things not supported by GitHub-hosted Windows runner Upgrade setuptools and wheel Rework windows test Restore root WORKSPACE to original Attribute naming fix in YML Enable one test in Windows Enable GPU Access Enable GPU Access Attempt 2 Enable GPU Access Attempt 3 - revert Switch off miniconda setup Specify Python 3.9 as default for testing Specify Python 3.9 as default for testing 2 Revert py39 changes, make changes to infra fork, re-enable cu121 installation Failing cuda installation Failing cuda installation 2 Failing cuda installation 3 Differentiate pre-build script during build and test Install trt bindings + lib manually Switch back to pytorch as repo source Specify validation runner Revert "Specify validation runner" This reverts commit 6d664c0. Enable more tests Enable more tests 2 Remove Python runtime specification Clean up and enable additional Windows tests Use .bat for driver upgrade Use .bat for driver upgrade 2 UTF8 encoding for windows testing
Remove serde and extraneous tests
b25d7c2
to
f8c300f
Compare
source "${BUILD_ENV_FILE}" | ||
|
||
# Install test index version of Torch and Torchvision | ||
${CONDA_RUN} pip install torch torchvision --index-url https://download.pytorch.org/whl/test/${CU_VERSION} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Actually you can restore this line back to ${CONDA_RUN} ${PIP_INSTALL_TORCH} torchvision
like before. The reason you had to manually specify test index is that, when you were using trt_10
as base branch, the set-channel
action set CHANNEL to nightly and hence PIP_INSTALL_TORCH was set to install nightly build. Now you have changed the base branch to release/2.3
, the set-channel
action set CHANNEL to test and hence PIP_INSTALL_TORCH was set to install test build.
python -m pip install -U numpy packaging pyyaml setuptools wheel | ||
|
||
# Install Torch from Test Index | ||
python -m pip install torch==2.3.0 --index-url https://download.pytorch.org/whl/test/${CU_VERSION} |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This line can be complete removed for the same reason as above, since the Install torch dependency step already installed the desired version.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Co-authored-by: HolyWu <holywu@gmail.com>
Co-authored-by: HolyWu <holywu@gmail.com>
Co-authored-by: HolyWu <holywu@gmail.com>
Co-authored-by: HolyWu <holywu@gmail.com>
Co-authored-by: HolyWu <holywu@gmail.com>
Co-authored-by: HolyWu <holywu@gmail.com>
Co-authored-by: HolyWu <holywu@gmail.com>
Description
Type of change
Checklist: