test: PyTorch PT2 Model Generation - #8666
Conversation
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
|
Label hould be |
|
Please fill out template for non-trivial PRs for easier tracking |
whoisj
changed the title
Updated. |
| @@ -0,0 +1,1000 @@ | |||
| tench | |||
There was a problem hiding this comment.
Why not curl download this file?
There was a problem hiding this comment.
Why "not"? Because that would just add overhead and an additional point of failure to the CI.
The file in only 11KB in the workspace, and less than 5KB in the repository.
There was a problem hiding this comment.
Pull request overview
Adds support for generating additional QA model artifacts using PyTorch PT2 AOTInductor (AOTI), alongside improved visibility/colorization of generation output to make long-running model generation less opaque in CI.
Changes:
- Add PT2/AOTI model generation + config generation paths to
qa/common/gen_qa_models.py(including TorchVision ResNet50 AOTI). - Add colorized/less-silent logging across the Python generator and the
gen_qa_model_repositorywrapper script. - Add a ResNet50 label file used by TorchVision/TorchTRT-style model configs; adjust model naming helper to allow a single-output model name format.
Reviewed changes
Copilot reviewed 4 out of 4 changed files in this pull request and generated 10 comments.
| File | Description |
|---|---|
| qa/common/test_util.py | Allow get_model_name() to omit output1_dtype for single-output model naming. |
| qa/common/resnet50_labels.txt | Add a 1000-class label list intended for ResNet50 output labeling. |
| qa/common/gen_qa_models.py | Introduce PT2/AOTI model+config generation and add colorized progress logging. |
| qa/common/gen_qa_model_repository | Improve logging/output visibility, add PT2/AOTI invocation, and adjust docker-run verbosity/behavior. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
This change adds the generation of PT2 format serialized PyTorch model files to the QA model generation scripts. add backend: pytorch to model config generation fix pedantic style complaints and very valid copyright mistakes pedantic style cop fix fix variable name error after rebase rename model add nvcc to pytorch path remove unnecessary changes add torch.export.export generation support to libtorch remove set -x from generated scripts improving gen_qa_models.py print statements fix color add print based on flags improve color visibility remove IDE help (caused errors) pedantic style cop fixes specify cuda fix libtorch pt2 generation add torchvision generation fix aoti model generation output types
Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com>
698eff3 to
f9fdaea
Compare
This change can be conceptually separated into three parts:
Commit Type:
Where should the reviewer start?
CI Pipeline ID:
44310672
Caveats:
Depends on nvcr.io/nvidia/pytorch:26.02-py3 and later containers.