Basic Bamba Rewrite + Numerical Precision #133

lucaslie · 2025-09-18T14:46:27Z

mostly based on this PR: NVIDIA#7745 by @2ez4bz

@coderabbitai summary

Description

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.

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.

Signed-off-by: William Zhang <133824995+2ez4bz@users.noreply.github.com>

Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com>

Copilot

Pull Request Overview

This PR implements a rewrite of the Bamba model patch to improve compatibility with torch.export and addresses numerical precision issues. It introduces custom operations for SSM transformations and provides comprehensive test coverage for the Bamba model export functionality.

Adds a complete Bamba model patch with custom SSM transformation operations
Implements test suite for verifying Bamba export functionality and numerical precision
Comments out a potential dynamic shape modification for attention interface

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 2 comments.

File	Description
tests/unittest/_torch/auto_deploy/unit/singlegpu/models/test_bamba.py	New comprehensive test file for Bamba model export with numerical precision validation
tensorrt_llm/_torch/auto_deploy/models/patches/bamba.py	Complete Bamba model patch implementation with custom SSM operations and export compatibility
tensorrt_llm/_torch/auto_deploy/custom_ops/attention_interface.py	Comments out dynamic shape modification line

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

Copilot · 2025-09-18T14:47:21Z

tensorrt_llm/_torch/auto_deploy/models/patches/bamba.py

+        module.dt_bias.data.fill_(1.0)
+        # ? Why is this even necessary? `BambaMixer.__init__` already sets
+        A = torch.arange(1, module.num_heads + 1)
+        self.A_log = nn.Parameter(torch.log(A))


Setting self.A_log on the parent class instance instead of the module parameter. This should be module.A_log = nn.Parameter(torch.log(A)) to properly initialize the parameter on the BambaMixer module.

Suggested change

self.A_log = nn.Parameter(torch.log(A))

module.A_log = nn.Parameter(torch.log(A))

Copilot · 2025-09-18T14:47:21Z

tensorrt_llm/_torch/auto_deploy/custom_ops/attention_interface.py

            if self.max_batch_size > 1:
                bs_seq_len_shape[0] = Dim("batch_size", max=self.max_batch_size)
            bs_seq_len_shape[1] = Dim("seq_len", max=self.max_seq_len)
+            # bs_seq_len_shape[1] = Dim.AUTO


Commented-out code should be removed rather than left in the codebase. If this alternative implementation is needed for future reference, consider adding a TODO comment explaining when it should be used.

Suggested change

# bs_seq_len_shape[1] = Dim.AUTO

* [None][auto_deploy] Bamba Signed-off-by: William Zhang <133824995+2ez4bz@users.noreply.github.com> * debugging export accuracy diff for bamba Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com> --------- Signed-off-by: William Zhang <133824995+2ez4bz@users.noreply.github.com> Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com> Co-authored-by: William Zhang <133824995+2ez4bz@users.noreply.github.com>

2ez4bz and others added 2 commits September 18, 2025 07:05

[None][auto_deploy] Bamba

1325c91

Signed-off-by: William Zhang <133824995+2ez4bz@users.noreply.github.com>

debugging export accuracy diff for bamba

267807f

Signed-off-by: Lucas Liebenwein <11156568+lucaslie@users.noreply.github.com>

Copilot AI review requested due to automatic review settings September 18, 2025 14:46

Copilot AI reviewed Sep 18, 2025

View reviewed changes

lucaslie merged commit bace21e into feat/ad_linear_attention Sep 18, 2025
2 of 5 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Basic Bamba Rewrite + Numerical Precision #133

Basic Bamba Rewrite + Numerical Precision #133

Uh oh!

lucaslie commented Sep 18, 2025 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Sep 18, 2025

Uh oh!

Copilot AI Sep 18, 2025

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	self.A_log = nn.Parameter(torch.log(A))
	module.A_log = nn.Parameter(torch.log(A))

Basic Bamba Rewrite + Numerical Precision #133

Basic Bamba Rewrite + Numerical Precision #133

Uh oh!

Conversation

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

Uh oh!

Description

Test Coverage

PR Checklist

GitHub Bot Help

kill

skip

reuse-pipeline

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Copilot AI Sep 18, 2025

Choose a reason for hiding this comment

Uh oh!

Copilot AI Sep 18, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

lucaslie commented Sep 18, 2025 •

edited

Loading