Could not import module 'Qwen3_5MoeForConditionalGeneration'

Checklist / 检查清单

• I have searched existing issues, and this is a new question or discussion topic. / 我已经搜索过现有的 issues，确认这是一个新的问题与讨论。

Question Description / 问题描述

环境配置如图：
ms_swift 4.0.0
transformer_engine 2.12.0
transformer_engine_cu12 2.12.0
transformer_engine_torch 2.12.0
transformers 5.2.0
transformers-stream-generator 0.0.5
推理qwen3.5模型报错：
/gemini/envs/qwen/lib/python3.10/site-packages/triton/runtime/jit.py", line 648, in init
src = src[re.search(r"^def\s+\w+\s*(", src, re.MULTILINE).start():]
AttributeError: 'NoneType' object has no attribute 'start'

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "/gemini/envs/qwen/demo.py", line 3, in
from transformers import AutoProcessor, Qwen3_5MoeForConditionalGeneration
File "/gemini/envs/qwen/lib/python3.10/site-packages/transformers/utils/import_utils.py", line 2132, in getattr
raise ModuleNotFoundError(
ModuleNotFoundError: Could not import module 'Qwen3_5MoeForConditionalGeneration'. Are this object's requirements defined correctly?

Start by reproducing the import in demo.py with the listed Python, transformers, ms-swift, and Transformer Engine versions. Trace the Qwen3_5MoeForConditionalGeneration import through the traceback and compare the failing dependency versions; done means the module imports and Qwen3.5 inference starts without the Triton error.