[https://nvbugs/6468821][fix] Rename the dynamic-tree worker's forward override to _forward_impl…

[chenfeiz0326]

Summary

• Root cause: MTPEagleDynamicTreeWorker overrode SpecWorkerBase.forward, which init_subclass forbids, raising TypeError at import and making import tensorrt_llm fail so the attention test collected 0 items.
• Fix: Rename the dynamic-tree worker's forward override to _forward_impl, matching the SpecWorkerBase contract and its parent MTPEagleWorker._forward_impl, so the base forward wrapper (with metadata cleanup) is preserved.
• Automated fix generated by repair-bot

Test plan

• Verify fix on the same GPU type as the original failure
• Check for regressions in related tests

Links

• Bug: https://nvbugs/6468821

Dev Engineer Review

• Renamed MTPEagleDynamicTreeWorker.forward(...) to _forward_impl(...) without changing its signature or implementation.
• Aligns the worker with the SpecWorkerBase contract and the existing MTPEagleWorker pattern, preventing the import-time TypeError while preserving base-wrapper metadata cleanup.
• No configuration, test-list, performance, or error-handling changes.

QA Engineer Review

No test changes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: cc9474db-1808-4f69-b45e-74372363e129

📥 Commits

Reviewing files that changed from the base of the PR and between 45d1c11 and 85ad978.

📒 Files selected for processing (1)

• tensorrt_llm/_torch/speculative/mtp_dynamic_tree.py

---

Walkthrough

MTPEagleDynamicTreeWorker.forward(...) is renamed to _forward_impl(...) without changing its signature, implementation body, or NVTX annotation.

Changes

Dynamic tree worker entry point

Layer / File(s)	Summary
Rename worker implementation method tensorrt_llm/_torch/speculative/mtp_dynamic_tree.py	Renames MTPEagleDynamicTreeWorker.forward(...) to _forward_impl(...) while preserving the method implementation and @nvtx_range("mtp_dyn.forward") annotation.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#16724: Makes the same method rename in the same file.
• NVIDIA/TensorRT-LLM#16740: Makes the same forward to _forward_impl rename.
• NVIDIA/TensorRT-LLM#16733: Includes the same rename alongside dynamic-tree attention state changes.

Suggested reviewers: zhaoyangwang-nvidia

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change and includes the bug ID and fix type.
Description check	✅ Passed	The description includes a clear summary, test plan, and bug link, so it is mostly complete despite not matching the template exactly.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[yuxianq]

The same fix #16724 has been merged, close this PR.