[https://nvbugs/5455836][fix] Fix llama 4 FP4 #6911
Conversation
📝 WalkthroughWalkthroughAdds a post-attention type normalization in Llama4Attention._forward_nope: if attn.forward returns a (tensor, scale) tuple, it is converted to an Fp4QuantizedTensor before passing into o_proj. No public APIs are changed. Changes
Sequence Diagram(s)sequenceDiagram
participant Caller
participant Llama4Attention
participant Attn
participant o_proj
Caller->>Llama4Attention: _forward_nope(q, k, v, ...)
Llama4Attention->>Attn: forward(q, k, v, ...)
Attn-->>Llama4Attention: attn_output (Tensor or (Tensor, scale))
alt attn_output is tuple
Llama4Attention->>Llama4Attention: wrap as Fp4QuantizedTensor(tensor, scale)
end
Llama4Attention->>o_proj: forward(attn_output)
o_proj-->>Llama4Attention: projected_output
Llama4Attention-->>Caller: projected_output
Estimated code review effort🎯 2 (Simple) | ⏱️ ~8 minutes Possibly related PRs
Suggested reviewers
Tip 🔌 Remote MCP (Model Context Protocol) integration is now available!Pro plan users can now connect to remote MCP servers from the Integrations page. Connect with popular remote MCPs such as Notion and Linear to add more context to your reviews and chats. ✨ Finishing Touches
🧪 Generate unit tests
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. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
|
/bot run |
|
PR_Github #15341 [ run ] triggered by Bot |
mikeiovine
changed the title
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com>
|
/bot run |
![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 (1)
tensorrt_llm/_torch/models/modeling_llama.py (1)
185-187: Guard against unexpected tuple arity and accept list for robustnessIf attn.forward ever changes to return more than two elements (or returns a list), this code will silently mis-handle the output. Add a small check and accept list as well to fail fast and aid debugging.
- if isinstance(attn_output, tuple): - attn_output = Fp4QuantizedTensor(attn_output[0], attn_output[1]) + # Normalize attention output to Fp4QuantizedTensor for o_proj. + if isinstance(attn_output, (tuple, list)): + if len(attn_output) != 2: + raise TypeError( + f"Unexpected attn_output length {len(attn_output)}; expected 2 (tensor, scale)" + ) + attn_output = Fp4QuantizedTensor(attn_output[0], attn_output[1])
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled
- Linear integration is disabled
You can enable these settings in your CodeRabbit configuration.
📒 Files selected for processing (1)
tensorrt_llm/_torch/models/modeling_llama.py(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+
Python indentation: 4 spaces, no tabs
Maintain module namespace in imports (from package.subpackage import foo; then use foo.SomeClass())
Python file names use snake_case
Python class names use PascalCase
Python functions/methods and local variables use snake_case; variables starting with a number get k_ prefix (e.g., k_99th_percentile)
Global variables use G_ prefixed UPPER_SNAKE_CASE (e.g., G_MY_GLOBAL)
Constants use UPPER_SNAKE_CASE in Python
Avoid shadowing variables from outer scopes in Python
Initialize all externally visible members of a Python class in init
Prefer docstrings for interfaces used outside a file; comments for local code
Use Google-style docstrings for classes and functions (Sphinx-parsable)
Document attributes/variables inline with short docstrings
Avoid reflection when simple alternatives exist (e.g., prefer explicit parameters over dict(**locals()))
In try/except, catch the narrowest exceptions possible
For duck-typing with try/except, keep try body minimal and put logic in else
Files:
tensorrt_llm/_torch/models/modeling_llama.py
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Prepend NVIDIA copyright header (current year) to all source files
Files:
tensorrt_llm/_torch/models/modeling_llama.py
⏰ 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 (2)
tensorrt_llm/_torch/models/modeling_llama.py (2)
185-187: Type normalization to Fp4QuantizedTensor before o_proj is correct and aligns with Linear’s contractConverting tuple outputs from attn.forward into Fp4QuantizedTensor here prevents tuple types from reaching o_proj and matches the intended API. This is consistent with the PR objective for llama 4 FP4.
185-187: No change needed — the RoPE path already normalizes attention outputs to Fp4QuantizedTensorShort explanation: Attention.forward converts (output, output_sf) -> Fp4QuantizedTensor before calling o_proj, so the super().forward (RoPE) path will pass a Fp4QuantizedTensor when required.
Places checked:
- tensorrt_llm/_torch/modules/attention.py — Attention.forward: after _attn_impl returns (output, output_sf) it does
if output_sf is not None: output = Fp4QuantizedTensor(output, output_sf)(≈ lines 460–464).- tensorrt_llm/_torch/models/modeling_llama.py — Llama4Attention._forward_nope already has the tuple->Fp4QuantizedTensor normalization (lines 185–186); this is consistent with the non-RoPE path.
- Note: the custom inplace op path (attn_custom_op_inplace / create_output) writes into a preallocated tensor in-place and does not return a tuple; create_output sets an appropriate dtype (e.g., float8) when needed, so no tuple wrapping is expected there.
Decision: no additional normalization needed in the RoPE (super().forward) path.
|
PR_Github #15343 [ run ] triggered by Bot |
|
PR_Github #15341 [ run ] completed with state |
|
PR_Github #15343 [ run ] completed with state |
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
Signed-off-by: Mike Iovine <6158008+mikeiovine@users.noreply.github.com> Signed-off-by: Wangshanshan <30051912+dominicshanshan@users.noreply.github.com>
5 participants
Description
Simple API issue was breaking llama 4. Fix was merged into main (#6690) but wasn't cherry picked into this release branch.
This PR is a bit different from #6690. Rather than checking for
tupleinsideLinear, I've made llama 4 attention always useFp4QuantizedTensor. This is more consistent with the rest of the system and the type annotations inLinear.Test Coverage
Existing tests
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.
Summary by CodeRabbit