[fix] Fix flaky mistral E2E test #6230
Conversation
WalkthroughThe test file for integration testing was updated by modifying the expected keywords in a specific test case for the "mistral-small-3.1-24b-instruct" model under the "image" modality. The keywords "stormy" and "rough" were replaced with "cloudy" and "water". No other changes were made. Changes
Estimated code review effort1 (~2 minutes) Suggested reviewers
Poem
📜 Recent review detailsConfiguration used: .coderabbit.yaml 📒 Files selected for processing (1)
✨ Finishing Touches
🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. Note: Be mindful of the bot's finite context window. It's strongly recommended to break down tasks such as reading entire modules into smaller chunks. For a focused discussion, use review comments to chat about specific files and their changes, instead of using the PR comments. CodeRabbit Commands (Invoked using PR comments)
Other keywords and placeholders
Documentation and Community
|
Signed-off-by: William Zhang <[email protected]>
2ez4bz
force-pushed
the
dev-fix-mistral-e2e-test
branch
from
994a2bf to
5e8cfe4
Compare
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🧹 Nitpick comments (3)
tensorrt_llm/_torch/models/checkpoints/hf/weight_loader.py (1)
27-29: Remove debug print statements before merging.The debug print statements should be removed as they're not appropriate for production code. Consider using proper logging if this information needs to be tracked.
Apply this diff to remove the debug prints:
- print(f"==== BEFORE: {len(weight_files)}") weight_files = [x for x in weight_files if "consolidated" not in x] - print(f"==== AFTER: {len(weight_files)}")Additionally, consider adding a comment explaining why consolidated files are filtered out, or use the logger instead:
+ # Filter out consolidated weight files as they may cause conflicts weight_files = [x for x in weight_files if "consolidated" not in x] + if len(weight_files) != len(orig_weight_files): + logger.info(f"Filtered out {len(orig_weight_files) - len(weight_files)} consolidated weight files")docker/Makefile (2)
157-158: Document the purpose of these volume mounts.The new volume mounts for scratch directories should be documented to explain their purpose and ensure they exist on target systems.
Consider adding comments to explain the purpose:
+ # Mount scratch directories for test data and GPU-specific caching --volume /home/scratch.trt_llm_data:/home/scratch.trt_llm_data \ --volume /home/scratch.williamz_gpu:/home/scratch.williamz_gpu \
164-165: Clean up commented code or document its purpose.The commented-out lines should either be removed if they're no longer needed, or uncommented if they're part of the intended changes.
Either remove the commented lines or add a comment explaining why they're temporarily disabled:
- # $(if $(filter 1,$(LOCAL_USER)),--volume ${HOME_DIR}/.cache:/home/${USER_NAME}/.cache:rw) \ - # --env TLLM_LLMAPI_BUILD_CACHE_ROOT=/home/scratch.williamz_gpu/trtllm_llmapi_cache \
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (4)
docker/Makefile(1 hunks)tensorrt_llm/_torch/models/checkpoints/hf/weight_loader.py(1 hunks)tensorrt_llm/builder.py(1 hunks)tests/integration/defs/test_e2e.py(1 hunks)
🔇 Additional comments (3)
docker/Makefile (1)
150-150: Consider security implications of host networking.Adding
--network=hostremoves network isolation between the container and host. While this may be necessary for certain testing scenarios, ensure this is intentional and documented.Please verify that host networking is required for the mistral E2E test fix and consider if a more restrictive networking approach could be used.
tests/integration/defs/test_e2e.py (1)
2079-2081: Good fix for the flaky test!This keyword update directly addresses the PR objective by making the test expectations align with the actual model output. Replacing "stormy" and "rough" with "cloudy" and "water" appears to be a reasonable adjustment based on observed model behavior.
The change successfully addresses the flakiness issue described in the PR by updating the expected keywords to match more consistent model outputs.
tensorrt_llm/builder.py (1)
688-694: LGTM! Robust type checking improves serialization safety.The explicit type checking for
kv_cache_typeis a good defensive programming practice that prevents silent serialization failures. The logic correctly handles both enum and string types while raising clear errors for unexpected types.Note that there's similar logic in the
ConfigEncoderclass (lines 51-53) that also handlesKVCacheTypeserialization. While this creates some redundancy, the explicit checking here provides an additional safety layer for theBuildConfig.to_dict()method.
|
@coderabbitai summary |
✅ Actions performedSummary regeneration triggered. |
|
/bot run |
|
PR_Github #12463 [ run ] triggered by Bot |
|
/bot cancel |
GitHub Bot Help
Provide a user friendly way for developers to interact with a Jenkins server. Run See details below for each supported subcommand.
Launch build/test pipelines. All previously running jobs will be killed.
kill
Kill all running builds associated with pull request. skip
Skip testing for latest commit on pull request. reuse-pipeline
Reuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break. |
|
/bot kill |
|
/bot run --extra-stage "H100_PCIe-PyTorch-Post-Merge-1" |
|
PR_Github #12464 [ kill ] triggered by Bot |
|
PR_Github #12463 [ run ] completed with state |
|
PR_Github #12464 [ kill ] completed with state |
|
PR_Github #12465 [ run ] triggered by Bot |
|
PR_Github #12465 [ run ] completed with state |
Signed-off-by: William Zhang <[email protected]> Signed-off-by: Shreyas Misra <[email protected]>
Signed-off-by: William Zhang <[email protected]> Signed-off-by: Ransiki Zhang <[email protected]>
Summary by CodeRabbit
[NVBugs 5404005][fix] Fix flaky mistral E2E test
Description
This PR updates the keywords checked in the mistral 3.1 small 24b E2E test.
Test Coverage
I was able to reproduce the flakiness by running the test function several times. Prior to this change, it failed 8 out of 10 times, whereas it passed all 10 after this change.
GitHub Bot Help
/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...Provide a user friendly way for developers to interact with a Jenkins server.
Run
/bot [-h|--help]to print this help message.See details below for each supported subcommand.
run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]Launch build/test pipelines. All previously running jobs will be killed.
--reuse-test (optional)pipeline-id(OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.--disable-reuse-test(OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.--disable-fail-fast(OPTIONAL) : Disable fail fast on build/tests/infra failures.--skip-test(OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.--stage-list "A10-PyTorch-1, xxx"(OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-1, xxx". Note: Does NOT update GitHub check status.--gpu-type "A30, H100_PCIe"(OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.--test-backend "pytorch, cpp"(OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline status.--only-multi-gpu-test(OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.--disable-multi-gpu-test(OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.--add-multi-gpu-test(OPTIONAL) : Force run the multi-GPU tests in addition to running L0 pre-merge pipeline.--post-merge(OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.--extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx"(OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".--detailed-log(OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.--debug(OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in thestage-listparameter to access the appropriate container environment. Note: Does NOT update GitHub check status.For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.mdand the
scripts/test_to_stage_mapping.pyhelper.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip testing for latest commit on pull request.
--comment "Reason for skipping build/test"is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.reuse-pipeline
reuse-pipelineReuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.