[https://nvbugs/5558117][fix] Allow per-layer quant config from hf_quant_config.json #8617
Conversation
📝 WalkthroughWalkthroughThis PR enhances quantization configuration handling by implementing a tolerant fallback mechanism in ModelConfig that attempts to load quant_cfg.json first, then falls back to hf_quant_config.json when unavailable. It aggregates configurations from both sources, validates kv_cache settings, builds per-layer QuantConfig entries, and ensures torch_dtype is consistently populated. Logging improvements truncate verbose output. Changes
Sequence DiagramsequenceDiagram
participant User
participant model_config as ModelConfig.load_<br/>modelopt_quant_config()
participant quant_cfg as quant_cfg.json
participant hf_quant as hf_quant_config.json
participant config as Merged Config
User->>model_config: load quantization config
model_config->>quant_cfg: try read quant_cfg.json
alt quant_cfg.json exists
quant_cfg-->>model_config: config data
model_config->>config: load quant_cfg
else quant_cfg.json missing/fails
model_config->>hf_quant: fallback: read hf_quant_config.json
hf_quant-->>model_config: config data
model_config->>model_config: log fallback message
model_config->>config: load hf_quant_config
end
rect rgb(200, 220, 240)
note over model_config,config: Aggregate & Validate
model_config->>config: merge both configs
model_config->>config: validate kv_cache_quant_algo<br/>compatibility across layers
end
model_config->>config: build per-layer QuantConfig<br/>with kv_cache_quant_algo
config-->>User: merged quantization config
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~20 minutes The changes involve new fallback logic and configuration aggregation with validation in the primary file, alongside straightforward logging truncation and two focused test cases. While the logic is relatively clear, the interaction between mixed quantization sources and per-layer config building requires careful verification of the aggregation and validation flow. Pre-merge checks and finishing touches❌ Failed checks (2 warnings)
✅ Passed checks (1 passed)
✨ Finishing touches
🧪 Generate unit tests (beta)
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: 4
Caution
Some comments are outside the diff and can’t be posted inline due to platform limitations.
⚠️ Outside diff range comments (2)
tensorrt_llm/llmapi/llm_utils.py (1)
1-1: Missing NVIDIA Apache-2.0 header (2025).Per coding guidelines, prepend the 2025 NVIDIA Apache-2.0 copyright header to all .py files.
Apply at file top:
+ # Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + # Licensed under the Apache License, Version 2.0 (the "License"); + # you may not use this file except in compliance with the License. + # You may obtain a copy of the License at + # http://www.apache.org/licenses/LICENSE-2.0 + # Unless required by applicable law or agreed to in writing, software + # distributed under the License is distributed on an "AS IS" BASIS, + # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + # See the License for the specific language governing permissions and + # limitations under the License.tensorrt_llm/_torch/model_config.py (1)
1-1: Missing NVIDIA Apache-2.0 header (2025).Add the standard 2025 NVIDIA Apache-2.0 header at the top.
+ # Copyright (c) 2025, NVIDIA CORPORATION. All rights reserved. + # Licensed under the Apache License, Version 2.0 (the "License"); + # you may not use this file except in compliance with the License. + # You may obtain a copy of the License at + # http://www.apache.org/licenses/LICENSE-2.0 + # Unless required by applicable law or agreed to in writing, software + # distributed under the License is distributed on an "AS IS" BASIS, + # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + # See the License for the specific language governing permissions and + # limitations under the License.
🧹 Nitpick comments (3)
tensorrt_llm/llmapi/llm_utils.py (1)
393-397: Avoid repeated str() and length checks in log.Compute once to reduce overhead and keep code clearer.
- for key, value in hf_quant_config.items(): - logger.info( - f"Setting {key}={str(value)[:100]}{'...' if len(str(value)) > 100 else ''} from HF quant config." - ) + for key, value in hf_quant_config.items(): + val_str = str(value) + truncated = (val_str[:100] + '...') if len(val_str) > 100 else val_str + logger.info(f"Setting {key}={truncated} from HF quant config.") setattr(quant_config, key, value)tests/unittest/llmapi/test_llm_quant.py (2)
79-137: Good coverage for quant_cfg.json + hf fallback; add a couple of assertions.Add asserts for layer kv_cache_quant_algo and use enum for consistency.
- assert quant_config.kv_cache_quant_algo == "FP8" + assert str(quant_config.kv_cache_quant_algo) == "FP8" + # Each per-layer config should inherit the global kv cache setting + assert layer_quant_config["model.layers.0.self_attn.k_proj"].kv_cache_quant_algo == "FP8"Optionally add a negative test for mismatched kv_cache between files raising RuntimeError.
139-174: Fallback-only path looks good; add one more per-layer kv cache check.Ensure per-layer kv cache inherits the global value in fallback case too.
- assert layer_quant_config[ - "model.layers.0.self_attn.q_proj"].quant_algo == "FP8" + assert layer_quant_config["model.layers.0.self_attn.q_proj"].quant_algo == "FP8" + assert layer_quant_config["model.layers.0.self_attn.q_proj"].kv_cache_quant_algo == "FP8"
📜 Review details
Configuration used: Path: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
tensorrt_llm/_torch/model_config.py(3 hunks)tensorrt_llm/llmapi/llm_utils.py(1 hunks)tests/unittest/llmapi/test_llm_quant.py(2 hunks)
🧰 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/llmapi/llm_utils.pytensorrt_llm/_torch/model_config.pytests/unittest/llmapi/test_llm_quant.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/llmapi/llm_utils.pytensorrt_llm/_torch/model_config.pytests/unittest/llmapi/test_llm_quant.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/llmapi/llm_utils.pytensorrt_llm/_torch/model_config.pytests/unittest/llmapi/test_llm_quant.py
🧬 Code graph analysis (2)
tensorrt_llm/_torch/model_config.py (3)
tensorrt_llm/llmapi/llm_args.py (4)
quant_config(2534-2537)quant_config(2540-2541)model_dir(1442-1444)model_dir(1447-1451)tensorrt_llm/models/modeling_utils.py (2)
QuantConfig(131-271)quant_algo(550-551)tensorrt_llm/logger.py (1)
info(138-139)
tests/unittest/llmapi/test_llm_quant.py (2)
tensorrt_llm/_torch/model_config.py (1)
load_modelopt_quant_config(258-321)tensorrt_llm/quantization/mode.py (1)
QuantAlgo(23-47)
🪛 Ruff (0.14.1)
tensorrt_llm/_torch/model_config.py
284-284: Do not catch blind exception: Exception
(BLE001)
286-286: f-string without any placeholders
Remove extraneous f prefix
(F541)
299-302: Avoid specifying long messages outside the exception class
(TRY003)
⏰ 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). (3)
- GitHub Check: Check PR Checklist Resolution
- GitHub Check: Check PR Title Format
- GitHub Check: Pre-commit Check
|
/bot run |
|
PR_Github #22261 [ run ] triggered by Bot. Commit: |
|
PR_Github #22261 [ run ] completed with state |
|
/bot run |
|
PR_Github #22501 [ run ] triggered by Bot. Commit: |
|
PR_Github #22501 [ run ] completed with state |
|
/bot run --disable-fail-fast |
|
PR_Github #22506 [ run ] triggered by Bot. Commit: |
|
PR_Github #22506 [ run ] completed with state |
|
/bot run Restart CI due to GB300s being offline |
|
PR_Github #22546 [ run ] triggered by Bot. Commit: |
|
PR_Github #22546 [ run ] completed with state |
rosenrodt
force-pushed
the
load-modelopt-perlayer-quant
branch
2 times, most recently
from
d17849f to
45f074b
Compare
|
/bot run --disable-fail-fast Retrigger CI due to B300 being offline |
|
PR_Github #22598 [ run ] triggered by Bot. Commit: |
|
PR_Github #22598 [ run ] completed with state |
rosenrodt
force-pushed
the
load-modelopt-perlayer-quant
branch
from
45f074b to
4cc93f0
Compare
|
/bot run Retrigger CI due to hitting HF rate limit in |
|
PR_Github #22655 [ run ] triggered by Bot. Commit: |
|
PR_Github #22655 [ run ] completed with state |
…cfg.json Signed-off-by: Anthony Chang <[email protected]>
rosenrodt
force-pushed
the
load-modelopt-perlayer-quant
branch
from
4cc93f0 to
a8e9206
Compare
|
/bot run |
|
PR_Github #22852 [ run ] triggered by Bot. Commit: |
|
PR_Github #22852 [ run ] completed with state |
|
/bot run --disable-reuse-test --disable-fail-fast |
|
PR_Github #22979 [ run ] triggered by Bot. Commit: |
|
/bot kill |
|
PR_Github #22983 [ kill ] triggered by Bot. Commit: |
|
PR_Github #22979 [ run ] completed with state |
|
PR_Github #22983 [ kill ] completed with state |
|
/bot reuse-pipeline Revert to last good pipeline from #8617 (comment). I accidentally triggered a new pipeline without making any change. |
|
PR_Github #23174 [ reuse-pipeline ] triggered by Bot. Commit: |
|
PR_Github #23174 [ reuse-pipeline ] completed with state |
Superjomn
left a comment
Superjomn
left a comment
There was a problem hiding this comment.
LGTM on the llmapi changes.
|
/bot skip --comment "pipeline passed before without code changes" |
|
PR_Github #23191 [ skip ] triggered by Bot. Commit: |
|
PR_Github #23191 [ skip ] completed with state |
…ant_config.json (NVIDIA#8617) Signed-off-by: Anthony Chang <[email protected]> Signed-off-by: FredricZ-2007 <[email protected]>
5 participants
Summary by CodeRabbit
Release Notes
New Features
Improvements
Tests
Description
This supports mixed quant recipe (e.g., FP8 QKV + W4A8 MoE) from ModelOpt PTQ models
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.