Save other CI jobs' result (torch/tf pipeline, example, deepspeed etc) #30699
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
I promise this is the last pr I am opening today 😆 |
|
well, this might need some extra work. No need to approve for now ...sorry being a bit too fast |
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
ydshieh
force-pushed
the
upload_artifacts_3
branch
from
6069a4d
to
d22c685
Compare
ydshieh
force-pushed
the
upload_artifacts_3
branch
from
d22c685
to
2ee86c4
Compare
ydshieh
commented
May 7, 2024
Comment on lines
+65
to
+66
| name: ci_results_${{ inputs.job }} | ||
| path: ci_results_${{ inputs.job }} |
There was a problem hiding this comment.
I tried not to change this again, but there doesn't seem a better solution.
The problem is that we can't upload the different artifacts to the same name (workflow run will fail).
ydshieh
commented
May 7, 2024
Comment on lines
+1151
to
+1162
| # Must have the same keys as in `additional_results`. | ||
| # The values are used as the file names where to save the corresponding CI job results. | ||
| test_to_result_name = { | ||
| "PyTorch pipelines": "torch_pipeline", | ||
| "TensorFlow pipelines": "tf_pipeline", | ||
| "Examples directory": "example", | ||
| "Torch CUDA extension tests": "deepspeed", | ||
| } | ||
| for job, job_result in additional_results.items(): | ||
| with open(f"ci_results_{job_name}/{test_to_result_name[job]}_results.json", "w", encoding="UTF-8") as fp: | ||
| json.dump(job_result, fp, indent=4, ensure_ascii=False) | ||
|
|
There was a problem hiding this comment.
This is the main changes in this PR.
ydshieh
commented
May 7, 2024
Comment on lines
+245
to
+250
| job_name = os.getenv("CI_TEST_JOB") | ||
| if not os.path.isdir(os.path.join(os.getcwd(), f"ci_results_{job_name}")): | ||
| os.makedirs(os.path.join(os.getcwd(), f"ci_results_{job_name}")) | ||
|
|
||
| with open(f"ci_results_{job_name}/quantization_results.json", "w", encoding="UTF-8") as fp: | ||
| json.dump(quantization_results, fp, indent=4, ensure_ascii=False) |
There was a problem hiding this comment.
Same as above but for quantization
ydshieh
commented
May 7, 2024
| @@ -416,7 +416,7 @@ def per_model_sum(model_category_dict): | |||
| reports=sorted_model_reports, | |||
| to_truncate=False, | |||
| ) | |||
| file_path = os.path.join(os.getcwd(), "ci_results/model_failures_report.txt") | |||
| file_path = os.path.join(os.getcwd(), f"ci_results_{job_name}/model_failures_report.txt") | |||
There was a problem hiding this comment.
(all same changes below too) the same reason as above
ydshieh
commented
May 7, 2024
Comment on lines
+87
to
+94
|
|
||
| # Upload complete failure tables, as they might be big and only truncated versions could be sent to Slack. | ||
| - name: Failure table artifacts | ||
| if: ${{ inputs.job == 'run_quantization_torch_gpu' }} | ||
| uses: actions/upload-artifact@v4 | ||
| with: | ||
| name: ci_results_${{ inputs.job }} | ||
| path: ci_results_${{ inputs.job }} |
There was a problem hiding this comment.
For quantization and only for it
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Currently we only save the model results. Let's save all of them.
I still need to work with the quantization job - as it is another separate notification file (and maybe also doctest - less priority for now).
I have to change the single artifact (
ci_results) to several one (ci_results_{job_name}) as we can't upload to the same name multiple times.See
https://github.com/huggingface/transformers/actions/runs/8988511794