Fix full-attention layer id mapping for Hybrid models (e.g. Qwen3-Next/Qwen3.5) - #249

Fix full-attention layer id mapping for Hybrid models (e.g. Qwen3-Next/Qwen3.5)#249

wanzhenchn wants to merge 1 commit into

Qwen3.5_v0.5.9zejunchen-zejun/sglang:Qwen3.5_v0.5.9from

fix-attention_layer_id_mappingzejunchen-zejun/sglang:fix-attention_layer_id_mappingCopy head branch name to clipboard