[#4674][bugfix] AutoDeploy Fix memory leak in fuse_moe #7844

galagam · 2025-09-18T16:52:55Z

Summary by CodeRabbit

New Features
- Enabled Mixture-of-Experts (MoE) fusion during post-load fusion, improving out-of-the-box optimization for supported models.
Performance
- Reduced peak GPU memory usage during MoE fusion by immediately removing dead subgraphs and unused components after fusion.
Tests
- Added unit tests to verify parameter reduction and stable CUDA memory after MoE fusion cleanup.

Description

AutoDeploy's fuse_moe transformation was previously disabled due to excessive memory usage. The transformation is stacking the expert weights, but fails to remove the unstacked weights that are no longer in use, causing a x2 memory bloating. This PR locates the unused parameters and removes them from the graph, allowing for memory deallocation.

Test Coverage

AutoDeploy dashboard doesn't crash with OOM.
_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py::test_fuse_moe_cleanup

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.

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 the stage-list parameter 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.md
and the scripts/test_to_stage_mapping.py helper.

kill

kill

Kill all running builds associated with pull request.

skip

skip --comment COMMENT

Skip 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-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.

lucaslie

have you tried your fix on large models, i.e., models with weights that take up >50% of available GPU memory and lots of experts?

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

galagam · 2025-09-22T08:22:18Z

have you tried your fix on large models, i.e., models with weights that take up >50% of available GPU memory and lots of experts?

@lucaslie Yes, this is only an issue for large models. Memory is cleaned after the transformation, so for small models this is not an issue. For large models, we crash during the transformation.

Please revisit, I modified the PR and marked as ready.

coderabbitai · 2025-09-22T08:32:26Z

📝 Walkthrough

Walkthrough

Enables MoE fusion in the default config. Updates the fused MoE transform to immediately eliminate dead code and unused submodules after node replacement. Adds a unit test verifying parameter node reduction and non-increasing CUDA memory after fusion with cleanup.

Changes

Cohort / File(s)	Summary
Config: Enable MoE fusion `tensorrt_llm/_torch/auto_deploy/config/default.yaml`	Set post_load_fusion.fuse_moe.enabled from false to true.
Transform: Fused MoE cleanup `tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py`	After MoE node replacement, invoke graph.eliminate_dead_code and GraphModule.delete_all_unused_submodules to remove dead subgraphs and unused submodules immediately.
Tests: MoE fusion cleanup `tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py`	Added test_fuse_moe_cleanup validating parameter node reduction and stable CUDA memory after fusion with cleanup.

Sequence Diagram(s)

sequenceDiagram
    autonumber
    participant C as Config
    participant IO as InferenceOptimizer
    participant FM as FusedMoE Transform
    participant G as GraphModule

    C->>IO: post_load_fusion.fuse_moe.enabled = true
    IO->>FM: Run fusion pass on G
    FM->>G: Replace MoE subgraph with fused node
    FM->>G: eliminate_dead_code()
    FM->>G: delete_all_unused_submodules()
    IO-->>C: Fusion complete with cleanup
    note over G: Reduced parameter nodes<br/>Unused modules removed

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Pre-merge checks and finishing touches

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	You can run `@coderabbitai generate docstrings` to improve docstring coverage.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Title Check	✅ Passed	The title "[#4674][bugfix] AutoDeploy Fix memory leak in fuse_moe" is concise, includes an issue reference and type, and accurately describes the primary change (fixing a memory leak in AutoDeploy's fuse_moe transformation), so it communicates the main intent to reviewers at a glance.
Description Check	✅ Passed	The PR description follows the repository template by providing a clear Description of the issue and fix, a Test Coverage section that references the unit test (tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py::test_fuse_moe_cleanup) and a dashboard check, and a PR Checklist; it explains the root cause (stacked expert weights not removing originals) and the remediation (removing unused parameters and cleanup). The required sections are present and sufficiently detailed to validate the change.

✨ Finishing touches

📝 Docstrings were successfully generated. (🔄 Check again to generate docstrings again)

🧪 Generate unit tests

Create PR with unit tests
Post copyable unit tests in a comment

📜 Recent review details

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

📥 Commits

Reviewing files that changed from the base of the PR and between 822cb01 and d8ee0ab.

📒 Files selected for processing (3)

tensorrt_llm/_torch/auto_deploy/config/default.yaml (1 hunks)
tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py (1 hunks)
tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py (1 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:

tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py
tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.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:

tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py
tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.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:

tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py
tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

🧠 Learnings (8)

📓 Common learnings

Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#6915
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:4010-4012
Timestamp: 2025-08-14T23:23:27.449Z
Learning: For MOE (Mixture of Experts) code reviews in TensorRT-LLM, avoid repeatedly suggesting finalize fusion validation checks and safety assertions. The user djns99 has indicated these suggestions are repetitive and unwanted across multiple MOE-related changes.

Learnt from: sklevtsov-nvidia
PR: NVIDIA/TensorRT-LLM#3294
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:1198-1209
Timestamp: 2025-08-08T22:03:40.707Z
Learning: In the CUTLASS MoE kernels (cpp/tensorrt_llm/cutlass_extensions), when `layout_info.fusion` is set to `TmaWarpSpecializedGroupedGemmInput::EpilogueFusion::FINALIZE`, the `router_scales` parameter must be non-null by design. The fused finalize kernel epilogue does not perform nullptr checks and requires valid router scales to function correctly. This is an implicit contract that callers must satisfy when enabling the FINALIZE fusion mode.

📚 Learning: 2025-08-14T23:23:27.449Z

Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#6915
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:4010-4012
Timestamp: 2025-08-14T23:23:27.449Z
Learning: For MOE (Mixture of Experts) code reviews in TensorRT-LLM, avoid repeatedly suggesting finalize fusion validation checks and safety assertions. The user djns99 has indicated these suggestions are repetitive and unwanted across multiple MOE-related changes.

Applied to files:

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

📚 Learning: 2025-08-21T02:39:12.009Z

Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#7104
File: cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu:1475-1480
Timestamp: 2025-08-21T02:39:12.009Z
Learning: The min latency mode functionality in TensorRT-LLM MOE kernels (cpp/tensorrt_llm/kernels/cutlass_kernels/moe_gemm/moe_kernels.cu) is deprecated and no longer being maintained/updated, as confirmed by djns99. Bug reports and optimization suggestions for the computeStridesTmaWarpSpecializedLowLatencyKernel and related min latency code paths should be deprioritized.

Applied to files:

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

📚 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/auto_deploy/transform/library/fused_moe.py

📚 Learning: 2025-08-22T01:54:35.850Z

Learnt from: djns99
PR: NVIDIA/TensorRT-LLM#7104
File: cpp/tensorrt_llm/kernels/cutlass_kernels/include/moe_kernels.h:999-1000
Timestamp: 2025-08-22T01:54:35.850Z
Learning: The `internal_cutlass_kernels` directory in TensorRT-LLM is a mirror of an internal NVIDIA repository and maintains its own implementation and API that may diverge from the public `cutlass_kernels` version. API inconsistencies between these two directories are intentional and by design, not bugs to be fixed.

Applied to files:

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

📚 Learning: 2025-08-11T20:09:24.389Z

Learnt from: achartier
PR: NVIDIA/TensorRT-LLM#6763
File: tests/integration/defs/triton_server/conftest.py:16-22
Timestamp: 2025-08-11T20:09:24.389Z
Learning: In the TensorRT-LLM test infrastructure, the team prefers simple, direct solutions (like hard-coding directory traversal counts) over more complex but robust approaches when dealing with stable directory structures. They accept the maintenance cost of updating tests if the layout changes.

Applied to files:

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

📚 Learning: 2025-07-22T09:22:14.726Z

Learnt from: yechank-nvidia
PR: NVIDIA/TensorRT-LLM#6254
File: tensorrt_llm/_torch/pyexecutor/model_engine.py:1201-1204
Timestamp: 2025-07-22T09:22:14.726Z
Learning: In TensorRT-LLM's multimodal processing pipeline, shared tensor recovery using `from_shared_tensor()` is only needed during the context phase. Generation requests reuse the already-recovered tensor data and only need to call `strip_for_generation()` to remove unnecessary multimodal data while preserving the recovered tensors. This avoids redundant tensor recovery operations during generation.

Applied to files:

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

📚 Learning: 2025-08-14T15:43:23.107Z

Learnt from: MatthiasKohl
PR: NVIDIA/TensorRT-LLM#6904
File: tensorrt_llm/_torch/attention_backend/trtllm.py:259-262
Timestamp: 2025-08-14T15:43:23.107Z
Learning: In TensorRT-LLM's attention backend, tensor parameters in the plan() method are assigned directly without validation (dtype, device, contiguity checks). This maintains consistency across all tensor inputs and follows the pattern of trusting callers to provide correctly formatted tensors.

Applied to files:

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py

🧬 Code graph analysis (1)

tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py (3)

tests/unittest/_torch/auto_deploy/_utils_test/_model_test_utils.py (2)

MoEOpModel (171-207)

get_input (206-207)

tensorrt_llm/_torch/auto_deploy/export/export.py (1)

torch_export_to_gm (198-273)

tensorrt_llm/_torch/auto_deploy/transform/optimizer.py (1)

InferenceOptimizer (22-84)

🔇 Additional comments (3)

tensorrt_llm/_torch/auto_deploy/config/default.yaml (1)

105-105: LGTM - Enables MoE fusion with memory leak fix in place.

This change enables the fuse_moe transformation which was previously disabled due to memory issues. The timing is appropriate since the memory leak fix is implemented in the corresponding transformation code.

tensorrt_llm/_torch/auto_deploy/transform/library/fused_moe.py (1)

60-65: Memory cleanup implementation looks correct.

The cleanup approach using gm.graph.eliminate_dead_code() and gm.delete_all_unused_submodules() immediately after node replacement is the right solution to prevent memory bloat during transformation. This addresses the core issue where unstacked weights remained in memory after stacking.

The inline comments clearly explain the rationale - preventing OOM during transformation rather than waiting for later canonicalization.

tests/unittest/_torch/auto_deploy/unit/singlegpu/transformations/library/test_moe_fusion.py (1)

373-418: Well-designed test validates the memory cleanup behavior.

The test effectively validates both aspects of the fix:

Parameter count reduction (ensuring unused parameters are removed)

Memory usage not increasing (ensuring cleanup is effective)

The test design is comprehensive:

Uses deterministic seeding for reproducibility

Explicit cache clearing and synchronization for accurate memory measurements

Disables other cleanup mechanisms to isolate the transformation's cleanup behavior

Clear assertion messages for debugging failures

The search confirms that both eliminate_dead_code() and delete_all_unused_submodules() are appropriate for cleaning up unused graph elements and reducing memory usage, as mentioned in item 1-11,1-12,1-16: the methods analyze the FX graph, identify unused submodules, and remove them to reduce memory footprint.

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.

Built-in checks – Quickly apply ready-made checks to enforce title conventions, require pull request descriptions that follow templates, validate linked issues for compliance, and more.
Custom agentic checks – Define your own rules using CodeRabbit’s advanced agentic capabilities to enforce organization-specific policies and workflows. For example, you can instruct CodeRabbit’s agent to verify that API documentation is updated whenever API schema files are modified in a PR. Note: Upto 5 custom checks are currently allowed during the preview period. Pricing for this feature will be announced in a few weeks.

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.

❤️ Share

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

galagam · 2025-09-22T08:33:35Z

@coderabbitai generate docstrings

coderabbitai · 2025-09-22T08:33:41Z

✅ Actions performed

Docstrings generation started.

coderabbitai · 2025-09-22T08:33:48Z

Caution

An unexpected error occurred while opening a pull request: Reference update failed - https://docs.github.com/rest/git/refs#create-a-reference

coderabbitai · 2025-09-22T08:34:18Z

Caution

An unexpected error occurred while opening a pull request: Reference update failed - https://docs.github.com/rest/git/refs#create-a-reference

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

galagam · 2025-09-28T08:10:31Z

/bot run

tensorrt-cicd · 2025-09-28T08:16:03Z

PR_Github #20162 [ run ] triggered by Bot

tensorrt-cicd · 2025-09-28T11:34:20Z

PR_Github #20162 [ run ] completed with state SUCCESS
/LLM/main/L0_MergeRequest_PR pipeline #15203 completed with status: 'SUCCESS'
Pipeline passed with automatic retried tests. Check the rerun report for details.

github-project-automation bot added this to AutoDeploy Board Sep 18, 2025

github-project-automation bot moved this to Backlog in AutoDeploy Board Sep 18, 2025

lucaslie reviewed Sep 18, 2025

View reviewed changes

galagam force-pushed the user/galagam/ad-fuse-moe-memleak branch from f598ea9 to 3c77148 Compare September 21, 2025 08:42

galagam marked this pull request as ready for review September 22, 2025 08:15

galagam requested a review from a team as a code owner September 22, 2025 08:15

galagam requested a review from lucaslie September 22, 2025 08:15

suyoggupta approved these changes Sep 26, 2025

View reviewed changes

galagam removed the request for review from lucaslie September 28, 2025 06:40

galagam added 8 commits September 28, 2025 09:41

[NVIDIA#4674][bugfix] AutoDeploy Fix memory leak in fuse_moe

524123b

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

simplify

28edfad

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

cleanup logging

38d7740

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

more cleanup

32174b4

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

more cleanup, prettify

c29e410

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

add unit test

2c3da09

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

simplify params delete

853dd28

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

use generic cleanup functions

d5aa8eb

Signed-off-by: Gal Hubara Agam <96368689+galagam@users.noreply.github.com>

galagam force-pushed the user/galagam/ad-fuse-moe-memleak branch from d8ee0ab to d5aa8eb Compare September 28, 2025 06:41

galagam merged commit b2095aa into NVIDIA:main Sep 29, 2025
5 checks passed

github-project-automation bot moved this from Backlog to Done in AutoDeploy Board Sep 29, 2025

galagam mentioned this pull request Oct 9, 2025

[AutoDeploy][Performance]: support MoE cuda-graph capture #7299

Closed

1 task

[#4674][bugfix] AutoDeploy Fix memory leak in fuse_moe #7844

[#4674][bugfix] AutoDeploy Fix memory leak in fuse_moe #7844

Uh oh!

Conversation

galagam commented Sep 18, 2025 • edited by coderabbitai bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary by CodeRabbit

Description

Test Coverage

PR Checklist

GitHub Bot Help

kill

skip

reuse-pipeline

Uh oh!

lucaslie left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

galagam commented Sep 22, 2025

Uh oh!

coderabbitai bot commented Sep 22, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Sequence Diagram(s)

Estimated code review effort

Pre-merge checks and finishing touches

Uh oh!

galagam commented Sep 22, 2025

Uh oh!

coderabbitai bot commented Sep 22, 2025

Uh oh!

coderabbitai bot commented Sep 22, 2025

Uh oh!

coderabbitai bot commented Sep 22, 2025

Uh oh!

galagam commented Sep 28, 2025

Uh oh!

tensorrt-cicd commented Sep 28, 2025

Uh oh!

tensorrt-cicd commented Sep 28, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

galagam commented Sep 18, 2025 •

edited by coderabbitai bot

Loading

coderabbitai bot commented Sep 22, 2025 •

edited

Loading