fix qwen3_next

[Jintao-Huang]

No description provided.

[gemini-code-assist]

Code Review

This pull request updates the forward method in qwen3_next.py to utilize resolve_gdn_attention_mask instead of resolve_hf_attention_mask. A review comment identifies potential logic errors and runtime risks within the resolve_gdn_attention_mask implementation, specifically regarding type safety when applying bitwise operators to integer tensors and potential dimensionality mismatches that could cause runtime errors.

[gemini-code-assist]

The function resolve_gdn_attention_mask (called here and defined at line 62) has a potential logic error and runtime risk:

1. Type Safety: At line 70, it uses the bitwise NOT operator ~ on attention_mask. If the mask is an integer tensor (common for HuggingFace masks), ~1 results in -2, which will cause the sum(...) > 0 check to behave unexpectedly. It should be explicitly cast to boolean: (~attention_mask.to(torch.bool)).
2. Dimensionality: The function assumes a 4D input by using dim=(1, 2). If kwargs.get('attention_mask') returns a 2D tensor (standard HuggingFace format), this will raise an IndexError.

Since this function is now being used to resolve the attention mask, these issues should be addressed in its implementation to ensure robustness.