[None][fix] Fix dummy load format for key models. #7993
Conversation
yuxianq
changed the title
|
/bot run --disable-fail-fast |
📝 WalkthroughWalkthroughRefactors post-load wiring across several Torch models by introducing or modifying post_load_weights hooks, relocates FP8 resmoothing to the linear module, toggles deep_gemm flags in Qwen3, adds an integration-test mode to AccuracyTask.evaluate, and expands test coverage with dummy load-format tests and test list entries. Removes an obsolete unit-test skip. Changes
Sequence Diagram(s)sequenceDiagram
autonumber
actor Loader as Weights Loader
participant Model as Model (e.g., Llama/GPT-OSS/Qwen3-MoE)
participant Layers as DecoderLayer[1..N]
participant LN as LayerNorm refs
participant Lin as Linear (FP8 QDQ)
Loader->>Model: load_state_dict(...)
note right of Model: Weights loaded
Model->>Model: post_load_weights()
loop For each layer i
Model->>Layers: wire input_layernorm / post_attention_layernorm
Model->>Layers: set next_layer_layernorm / next_attn refs
end
alt FP8 QDQ present
Model->>Lin: FP8QDQLinearMethod.post_load_weights(module)
Lin->>Lin: resmooth_to_fp8_e8m0(...)
Lin->>Lin: transform_sf_into_required_layout(...)
Lin-->>Model: update weight and weight_scale
end
sequenceDiagram
autonumber
actor Test as Test Runner
participant Task as AccuracyTask
participant LLM as LLM (dummy or real)
Test->>LLM: construct(load_format="dummy") [optional]
Test->>Task: evaluate(..., is_integration_test=True/False)
alt Integration mode
Task->>Task: num_samples=1, threshold=0
Task-->>Test: run minimal check
else Normal mode
Task->>Task: compute num_samples, threshold from spec
Task-->>Test: run full accuracy verification
end
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 passed)
✨ Finishing touches
🧪 Generate unit tests
Tip 👮 Agentic pre-merge checks are now available in preview!Pro plan users can now enable pre-merge checks in their settings to enforce checklists before merging PRs.
Please see the documentation for more information. Example: reviews:
pre_merge_checks:
custom_checks:
- name: "Undocumented Breaking Changes"
mode: "warning"
instructions: |
Pass/fail criteria: All breaking changes to public APIs, CLI flags, environment variables, configuration keys, database schemas, or HTTP/GraphQL endpoints must be documented in the "Breaking Change" section of the PR description and in CHANGELOG.md. Exclude purely internal or private changes (e.g., code not exported from package entry points or explicitly marked as internal).Please share your feedback with us on this Discord post. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (3)
tensorrt_llm/_torch/models/modeling_qwen3.py (2)
54-69: Re-enabling deep_gemm for Qwen3 may reintroduce known accuracy issues; consider gating or making opt-in.Comments above reference open accuracy bugs; flipping
disable_deep_gemm = Falseunconditionally risks regressions. Suggest gating by SM version, quant mode, or an env/config toggle (e.g., enable only on Blackwell or when explicitly requested).Example (conceptual):
- Keep disabled unless
get_sm_version() >= 100and an override flagTRTLLM_ENABLE_DEEP_GEMM_QWEN3=1is set.
91-101: Same concern for MLP path: deep_gemm re-enabled unconditionally.Apply the same gating/toggle strategy here to avoid MLP-side accuracy regressions.
tensorrt_llm/_torch/modules/linear.py (1)
720-735: Device safety and dtype/layout verification for FP8 resmoothing.
- resmooth helper uses
.cuda()internally; in multi‑GPU, ensure tensors land on the same device asmodule.weightto avoid cross‑device issues.- After replacement,
module.weightlikely becomes E8M0 (per helper), which is expected byfp8_swap_ab_gemm. Confirm this aligns with all call paths whenis_sm_100f()is true.Suggested patch to ensure device consistency:
if is_sm_100f(): - weight, weight_scale = resmooth_to_fp8_e8m0(module.weight, - module.weight_scale) + weight, weight_scale = resmooth_to_fp8_e8m0(module.weight, module.weight_scale) + # Ensure tensors are on the original module device + dev = module.weight.device + weight = weight.to(dev) + weight_scale = weight_scale.to(dev) transfromed_scale = transform_sf_into_required_layout( weight_scale, mn=weight.shape[0], k=weight.shape[1], recipe=(1, 128, 128), is_sfa=False) module.weight = nn.Parameter(weight, requires_grad=False) module.weight_scale = nn.Parameter( transfromed_scale, requires_grad=False, )Also please verify kernels expect E8M0 weights + INT packed scales under this path.
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (11)
tensorrt_llm/_torch/models/modeling_deepseekv3.py(1 hunks)tensorrt_llm/_torch/models/modeling_gpt_oss.py(1 hunks)tensorrt_llm/_torch/models/modeling_llama.py(2 hunks)tensorrt_llm/_torch/models/modeling_qwen3.py(2 hunks)tensorrt_llm/_torch/models/modeling_qwen3_moe.py(1 hunks)tensorrt_llm/_torch/modules/linear.py(2 hunks)tests/integration/defs/accuracy/accuracy_core.py(2 hunks)tests/integration/defs/accuracy/test_llm_api_pytorch.py(5 hunks)tests/integration/test_lists/test-db/l0_b200.yml(1 hunks)tests/integration/test_lists/test-db/l0_h100.yml(2 hunks)tests/unittest/_torch/modeling/test_modeling_gpt_oss.py(0 hunks)
💤 Files with no reviewable changes (1)
- tests/unittest/_torch/modeling/test_modeling_gpt_oss.py
🧰 Additional context used
📓 Path-based instructions (3)
**/*.{h,hpp,hh,hxx,cpp,cxx,cc,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Use only spaces, no tabs; indent with 4 spaces.
Files:
tensorrt_llm/_torch/models/modeling_deepseekv3.pytensorrt_llm/_torch/models/modeling_qwen3.pytensorrt_llm/_torch/models/modeling_qwen3_moe.pytests/integration/defs/accuracy/test_llm_api_pytorch.pytensorrt_llm/_torch/modules/linear.pytests/integration/defs/accuracy/accuracy_core.pytensorrt_llm/_torch/models/modeling_gpt_oss.pytensorrt_llm/_torch/models/modeling_llama.py
**/*.py
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+.
Indent Python code with 4 spaces; do not use tabs.
Maintain module namespace when importing; prefer 'from package.subpackage import foo' then 'foo.SomeClass()' instead of importing the class directly.
Python filenames should be snake_case (e.g., some_file.py).
Python classes use PascalCase names.
Functions and methods use snake_case names.
Local variables use snake_case; prefix 'k' for variables that start with a number (e.g., k_99th_percentile).
Global variables use upper SNAKE_CASE prefixed with 'G' (e.g., G_MY_GLOBAL).
Constants use upper SNAKE_CASE (e.g., MY_CONSTANT).
Avoid shadowing variables from an outer scope.
Initialize all externally visible members of a class in the constructor.
Prefer docstrings for interfaces that may be used outside a file; comments for in-function or file-local interfaces.
Use Google-style docstrings for classes and functions (Sphinx-parsable).
Document attributes and variables inline so they render under the class/function docstring.
Avoid reflection when a simpler, explicit approach suffices (e.g., avoid dict(**locals()) patterns).
In try/except, catch the most specific exceptions possible.
For duck-typing try/except, keep the try body minimal and use else for the main logic.
Files:
tensorrt_llm/_torch/models/modeling_deepseekv3.pytensorrt_llm/_torch/models/modeling_qwen3.pytensorrt_llm/_torch/models/modeling_qwen3_moe.pytests/integration/defs/accuracy/test_llm_api_pytorch.pytensorrt_llm/_torch/modules/linear.pytests/integration/defs/accuracy/accuracy_core.pytensorrt_llm/_torch/models/modeling_gpt_oss.pytensorrt_llm/_torch/models/modeling_llama.py
**/*.{cpp,cxx,cc,h,hpp,hh,hxx,cu,cuh,py}
📄 CodeRabbit inference engine (CODING_GUIDELINES.md)
Prepend the NVIDIA Apache-2.0 copyright header with current year to the top of all source files (e.g., .cpp, .h, .cu, .py).
Files:
tensorrt_llm/_torch/models/modeling_deepseekv3.pytensorrt_llm/_torch/models/modeling_qwen3.pytensorrt_llm/_torch/models/modeling_qwen3_moe.pytests/integration/defs/accuracy/test_llm_api_pytorch.pytensorrt_llm/_torch/modules/linear.pytests/integration/defs/accuracy/accuracy_core.pytensorrt_llm/_torch/models/modeling_gpt_oss.pytensorrt_llm/_torch/models/modeling_llama.py
🧠 Learnings (5)
📚 Learning: 2025-08-19T12:45:11.997Z
Learnt from: amitz-nv
PR: NVIDIA/TensorRT-LLM#7033
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:0-0
Timestamp: 2025-08-19T12:45:11.997Z
Learning: In tensorrt_llm/_torch/pyexecutor/model_engine.py, DoRA (Delta Orthogonal Rank Adaptation) functionality was removed from the PyTorch flow to eliminate issues with inverted DoRA detection logic. The original is_dora condition was checking if scaling_vec_pointer == 0, which was potentially incorrect.
Applied to files:
tensorrt_llm/_torch/models/modeling_deepseekv3.py
📚 Learning: 2025-09-09T09:40:45.658Z
Learnt from: fredricz-20070104
PR: NVIDIA/TensorRT-LLM#7645
File: tests/integration/test_lists/qa/llm_function_core.txt:648-648
Timestamp: 2025-09-09T09:40:45.658Z
Learning: In TensorRT-LLM test lists, it's common and intentional for the same test to appear in multiple test list files when they serve different purposes (e.g., llm_function_core.txt for comprehensive core functionality testing and llm_function_core_sanity.txt for quick sanity checks). This duplication allows tests to be run in different testing contexts.
Applied to files:
tests/integration/test_lists/test-db/l0_h100.ymltests/integration/test_lists/test-db/l0_b200.yml
📚 Learning: 2025-09-17T02:48:52.732Z
Learnt from: tongyuantongyu
PR: NVIDIA/TensorRT-LLM#7781
File: tests/integration/test_lists/waives.txt:313-313
Timestamp: 2025-09-17T02:48:52.732Z
Learning: In TensorRT-LLM, `tests/integration/test_lists/waives.txt` is specifically for waiving/skipping tests, while other test list files like those in `test-db/` and `qa/` directories are for different test execution contexts (pre-merge, post-merge, QA tests). The same test appearing in both waives.txt and execution list files is intentional - the test is part of test suites but will be skipped due to the waiver.
Applied to files:
tests/integration/test_lists/test-db/l0_h100.yml
📚 Learning: 2025-07-28T17:06:08.621Z
Learnt from: moraxu
PR: NVIDIA/TensorRT-LLM#6303
File: tests/integration/test_lists/qa/examples_test_list.txt:494-494
Timestamp: 2025-07-28T17:06:08.621Z
Learning: In TensorRT-LLM testing, it's common to have both CLI flow tests (test_cli_flow.py) and PyTorch API tests (test_llm_api_pytorch.py) for the same model. These serve different purposes: CLI flow tests validate the traditional command-line workflow, while PyTorch API tests validate the newer LLM API backend. Both are legitimate and should coexist.
Applied to files:
tests/integration/test_lists/test-db/l0_h100.ymltests/integration/defs/accuracy/test_llm_api_pytorch.pytests/integration/test_lists/test-db/l0_b200.yml
📚 Learning: 2025-09-03T13:16:06.824Z
Learnt from: nvpohanh
PR: NVIDIA/TensorRT-LLM#7478
File: tensorrt_llm/_torch/models/modeling_llama.py:1315-1315
Timestamp: 2025-09-03T13:16:06.824Z
Learning: The Llama4VisionEncoder.load_weights method signature is `def load_weights(self, weights: Dict)` and should not be confused with Llama4ForConditionalGeneration.load_weights which has a different signature including weight_mapper parameter.
Applied to files:
tensorrt_llm/_torch/models/modeling_qwen3_moe.pytensorrt_llm/_torch/models/modeling_llama.py
🧬 Code graph analysis (6)
tensorrt_llm/_torch/models/modeling_deepseekv3.py (1)
tensorrt_llm/_utils.py (1)
get_sm_version(689-691)
tensorrt_llm/_torch/models/modeling_qwen3_moe.py (10)
tensorrt_llm/_torch/models/modeling_deepseekv3.py (1)
post_load_weights(1520-1527)tensorrt_llm/_torch/models/modeling_gpt_oss.py (1)
post_load_weights(603-614)tensorrt_llm/_torch/models/modeling_llama.py (2)
post_load_weights(986-994)post_load_weights(1312-1320)tensorrt_llm/_torch/modules/linear.py (3)
post_load_weights(246-247)post_load_weights(720-734)post_load_weights(2026-2027)tensorrt_llm/_torch/modules/fused_moe/fused_moe_wide_ep.py (1)
post_load_weights(1039-1040)tensorrt_llm/_torch/modules/fused_moe/fused_moe_cutlass.py (1)
post_load_weights(602-603)tensorrt_llm/_torch/modules/fused_moe/fused_moe_trtllm_gen.py (1)
post_load_weights(172-173)tensorrt_llm/_torch/modules/fused_moe/interface.py (1)
post_load_weights(198-199)tensorrt_llm/_torch/modules/fused_moe/fused_moe_triton.py (1)
post_load_weights(1389-1390)tensorrt_llm/_torch/modules/fused_moe/quantization.py (2)
post_load_weights(325-326)post_load_weights(732-752)
tests/integration/defs/accuracy/test_llm_api_pytorch.py (4)
tensorrt_llm/llmapi/llm.py (1)
LLM(1052-1068)tests/integration/defs/accuracy/accuracy_core.py (4)
MMLU(278-292)evaluate(147-208)evaluate(714-724)GSM8K(295-310)tests/integration/defs/conftest.py (2)
llm_models_root(77-91)get_sm_version(1868-1871)tensorrt_llm/llmapi/llm_args.py (1)
MoeConfig(168-196)
tensorrt_llm/_torch/modules/linear.py (2)
tensorrt_llm/quantization/utils/fp8_utils.py (2)
resmooth_to_fp8_e8m0(82-92)transform_sf_into_required_layout(169-217)tensorrt_llm/_utils.py (3)
is_sm_100f(695-698)shape(955-956)shape(972-973)
tensorrt_llm/_torch/models/modeling_gpt_oss.py (3)
tensorrt_llm/_torch/models/modeling_deepseekv3.py (1)
post_load_weights(1520-1527)tensorrt_llm/_torch/models/modeling_llama.py (2)
post_load_weights(986-994)post_load_weights(1312-1320)tensorrt_llm/_torch/models/modeling_qwen3_moe.py (1)
post_load_weights(391-398)
tensorrt_llm/_torch/models/modeling_llama.py (4)
tensorrt_llm/_torch/models/modeling_deepseekv3.py (1)
post_load_weights(1520-1527)tensorrt_llm/_torch/models/modeling_gpt_oss.py (1)
post_load_weights(603-614)tensorrt_llm/_torch/models/modeling_qwen3_moe.py (1)
post_load_weights(391-398)tensorrt_llm/_torch/models/modeling_llama_min_latency.py (2)
post_load_weights(310-317)post_load_weights(566-580)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Pre-commit Check
🔇 Additional comments (17)
tensorrt_llm/_torch/models/modeling_deepseekv3.py (1)
43-43: LGTM: import slimming to only get get_sm_version.No functional risk; consistent with current usage in file.
tensorrt_llm/_torch/models/modeling_qwen3_moe.py (1)
391-399: Post-load next_layer_layernorm wiring: OK.Matches the pattern used across models; no issues spotted.
tensorrt_llm/_torch/modules/linear.py (1)
23-25: FP8 utilities import: OK.Imports are scoped and used by post-load smoothing; no concerns.
tensorrt_llm/_torch/models/modeling_llama.py (2)
986-995: post_load_weights wiring for Llama: OK.Correctly assigns next layernorm and next attention references.
1312-1321: Llama4 post-load wiring: OK.Consistent with Llama; no issues spotted.
tensorrt_llm/_torch/models/modeling_gpt_oss.py (1)
603-615: post_load_weights invocation confirmed model_engine.py iterates over all modules after loading weights and callspost_load_weights(), covering the GPT-OSS model.tests/integration/test_lists/test-db/l0_b200.yml (2)
37-37: LGTM! Test entry properly added for DeepSeekV3Lite.The new test entry
test_dummy_load_formatis correctly positioned after the guided decoding test and follows the established test naming pattern.
42-42: LGTM! Test entry properly added for GPTOSS.The new test entry
test_dummy_load_formatis correctly positioned after the w4_1gpu test variants and maintains consistency with the test suite structure.tests/integration/test_lists/test-db/l0_h100.yml (4)
46-46: LGTM! Test entry correctly added for Llama3_1_8BInstruct.The new test entry
test_dummy_load_formatis properly positioned after the chunked_prefill test and follows the established test naming convention.
64-64: LGTM! Test entry properly added for DeepSeekV3Lite.The new test entry is correctly positioned between the chunked_prefill variants, maintaining the test organization structure.
66-66: LGTM! Test entry properly added for Qwen3_8B.The new test entry is appropriately placed after the fp8_block_scales test and before the Qwen3_30B_A3B tests.
69-69: LGTM! Test entry properly added for Qwen3_30B_A3B.The new test entry is correctly positioned between the fp8 test variants and maintains consistency with the test suite organization.
tests/integration/defs/accuracy/test_llm_api_pytorch.py (5)
90-96: LGTM! Test method properly implements dummy load format validation for Llama.The test correctly initializes an LLM with
load_format="dummy"and evaluates the MMLU task withis_integration_test=Trueto skip accuracy verification.
1904-1915: LGTM! Test method correctly implements dummy load format validation for DeepSeekV3Lite.The test properly configures MoE backend selection based on SM version (using DEEPGEMM for SM100+ or CUTLASS otherwise) and evaluates the GSM8K task in integration test mode.
2624-2633: LGTM! Test method properly implements dummy load format validation for Qwen3_8B.The test correctly uses FP8 model path with dummy load format and evaluates MMLU task with integration test flag.
2752-2761: LGTM! Test method properly implements dummy load format validation for Qwen3_30B_A3B.The test correctly uses FP8 model path with dummy load format and evaluates GSM8K task with integration test flag.
3266-3275: LGTM! Test method properly implements dummy load format validation for GPTOSS.The test correctly uses the 20B model variant with dummy load format and evaluates GSM8K task in integration test mode.
|
PR_Github #19939 [ run ] triggered by Bot |
|
PR_Github #19939 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #20084 [ run ] triggered by Bot |
|
/bot run --disable-fail-fast |
|
PR_Github #20086 [ run ] triggered by Bot |
|
PR_Github #20084 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #20088 [ run ] triggered by Bot |
|
PR_Github #20086 [ run ] completed with state |
|
PR_Github #20088 [ run ] completed with state |
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
yuxianq
force-pushed
the
fix-dummy-load
branch
from
42e44ea to
b0d8ea1
Compare
|
/bot run --disable-fail-fast |
|
PR_Github #20258 [ run ] triggered by Bot |
|
PR_Github #20258 [ run ] completed with state |
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
|
/bot run --disable-fail-fast |
|
PR_Github #20358 [ run ] triggered by Bot |
|
PR_Github #20358 [ run ] completed with state |
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com> Signed-off-by: yufeiwu-nv <230315618+yufeiwu-nv@users.noreply.github.com>
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
Signed-off-by: Yuxian Qiu <142763828+yuxianq@users.noreply.github.com>
…lclose_to_hf failure This test was fixed in NVIDIA#7478 but was broken by NVIDIA#7993 because the latter PR moved the next_layer_layernorm setting logic from load_weights() to post_load_weights() but the test definition was not updated. The failure was hidden because we did not upgrade transformers version yet. To fix this, remember to call post_load_weights() after load_weights() in test definition (post_load_weights() is called automatically if the users are using pyexecutor). Also, raise a runtime error if the users forget to do so to avoid silent corrupted output which is difficult to debug. Signed-off-by: Po-Han Huang <pohanh@nvidia.com>
…lclose_to_hf failure This test was fixed in NVIDIA#7478 but was broken by NVIDIA#7993 because the latter PR moved the next_layer_layernorm setting logic from load_weights() to post_load_weights() but the test definition was not updated. The failure was hidden because we did not upgrade transformers version yet. To fix this, remember to call post_load_weights() after load_weights() in test definition (post_load_weights() is called automatically if the users are using pyexecutor). Also, raise a runtime error if the users forget to do so to avoid silent corrupted output which is difficult to debug. Signed-off-by: Po-Han Huang <pohanh@nvidia.com>
…lclose_to_hf failure This test was fixed in NVIDIA#7478 but was broken by NVIDIA#7993 because the latter PR moved the next_layer_layernorm setting logic from load_weights() to post_load_weights() but the test definition was not updated. The failure was hidden because we did not upgrade transformers version yet. To fix this, remember to call post_load_weights() after load_weights() in test definition (post_load_weights() is called automatically if the users are using pyexecutor). Also, raise a runtime error if the users forget to do so to avoid silent corrupted output which is difficult to debug. Signed-off-by: Po-Han Huang <pohanh@nvidia.com>
5 participants
Summary by CodeRabbit
New Features
Performance
Bug Fixes
Tests
Description
This PR fix dummy load format for Qwen/LLaMA/GPT-OSS and add tests for these cases.
Test Coverage
PR Checklist
Please review the following before submitting your PR:
PR description clearly explains what and why. If using CodeRabbit's summary, please make sure it makes sense.
PR Follows TRT-LLM CODING GUIDELINES to the best of your knowledge.
Test cases are provided for new code paths (see test instructions)
Any new dependencies have been scanned for license and vulnerabilities
CODEOWNERS updated if ownership changes
Documentation updated as needed
The reviewers assigned automatically/manually are appropriate for the PR.
Please check this after reviewing the above items as appropriate for this PR.
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.
Details
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.