[model] fix: improve apply_rope_fusion assert message for Qwen3-VL

[yaoyu-33]

Summary

• Replaces bare assert not config.apply_rope_fusion in rope.py with a descriptive assertion message and detailed comments explaining why fused RoPE is unsupported for Qwen3-VL / Qwen3.5-VL
• Verified on cluster that even removing the assert and forcing apply_rope_fusion=True never actually enables TE fused kernels due to three upstream guards:
  1. validate_rope_fusion_compatibility() blocks position_embedding_type="mrope"
  2. Qwen3VLSelfAttention calls apply_rotary_pos_emb_absolute() directly (bypasses fused dispatch)
  3. Per-token absolute freq format (seq_len, bs, 1, 2*dim) is incompatible with TE's expected (max_seqlen, 1, 1, 2*dim)

Closes #3296

Test plan

• Existing unit tests pass (tests/unit_tests/models/qwen_vl/)
• No functional change — assert behavior is identical, only the error message improves

🤖 Generated with Claude Code

Summary by CodeRabbit

• Bug Fixes
  • Enhanced configuration validation error messages to help users better understand which settings are incompatible and why.

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[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: Pro Plus

Run ID: 808e89a9-9e93-4c6c-a1ee-e1a59e4ede6f

📥 Commits

Reviewing files that changed from the base of the PR and between 9b38f65 and d91435c.

📒 Files selected for processing (1)

• src/megatron/bridge/models/qwen_vl/modelling_qwen3_vl/rope.py

---

📝 Walkthrough

Walkthrough

A bare assertion preventing apply_rope_fusion configuration in Qwen3-VL and Qwen3.5-VL models was replaced with a guarded assertion that provides a detailed error message explaining why the feature must remain disabled.

Changes

Cohort / File(s)	Summary
Error Message Enhancement src/megatron/bridge/models/qwen_vl/modelling_qwen3_vl/rope.py	Added explicit, descriptive error message to assertion that validates apply_rope_fusion is disabled for Qwen3-VL and Qwen3.5-VL models, replacing bare assert statement.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

🚥 Pre-merge checks | ✅ 6

✅ Passed checks (6 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly summarizes the main change: improving the assert message for apply_rope_fusion in Qwen3-VL, which is exactly what the changeset accomplishes.
Linked Issues check	✅ Passed	The PR adequately addresses issue #3296 by replacing the bare assert with a descriptive message and detailed comments explaining why apply_rope_fusion is unsupported, providing the clarification requested.
Out of Scope Changes check	✅ Passed	All changes are scoped to improving the assertion message and documentation in rope.py as specified by the linked issue; no unrelated changes are present.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Test Results For Major Changes	✅ Passed	Minor change improving assertion error message with no functional impact; existing unit tests documented to pass.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/qwen3vl-rope-fusion-assert-3296

---

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