使用vulkan后端进行qwen1.5b的推理时，报Don't support type [Attention], Attention/Reshape_8_output_0错误

[RisingEntropy]

我在windows上用2.9.6版本的MNN进行qwen 1.5b的推理，后端选择vulkan时报错如下，使用cpu后端一切正常：

The device supports: i8sdot:0, fp16:0, i8mm: 0, sve2: 0
Can't open file:.tempcache
Load Cache file error.
### is_single_ = 1
load tokenizer
tokenizer_type = 3
load tokenizer Done
load D:\llama2\model\llm.mnn ... Don't support type [Attention], Attention/Reshape_8_output_0
Don't support type [Attention], Attention/Reshape_8_output_0
Don't support type [Attention], Attention/Reshape_17_output_0
Don't support type [Attention], Attention/Reshape_17_output_0
Don't support type [Attention], Attention/Reshape_26_output_0
Don't support type [Attention], Attention/Reshape_26_output_0
Don't support type [Attention], Attention/Reshape_35_output_0
Don't support type [Attention], Attention/Reshape_35_output_0
Don't support type [Attention], Attention/Reshape_44_output_0
Don't support type [Attention], Attention/Reshape_44_output_0
Don't support type [Attention], Attention/Reshape_53_output_0
Don't support type [Attention], Attention/Reshape_53_output_0
Don't support type [Attention], Attention/Reshape_62_output_0
Don't support type [Attention], Attention/Reshape_62_output_0
Don't support type [Attention], Attention/Reshape_71_output_0
Don't support type [Attention], Attention/Reshape_71_output_0
Don't support type [Attention], Attention/Reshape_80_output_0
Don't support type [Attention], Attention/Reshape_80_output_0
Don't support type [Attention], Attention/Reshape_89_output_0
Don't support type [Attention], Attention/Reshape_89_output_0
Don't support type [Attention], Attention/Reshape_98_output_0
Don't support type [Attention], Attention/Reshape_98_output_0
Don't support type [Attention], Attention/Reshape_107_output_0
Don't support type [Attention], Attention/Reshape_107_output_0
Don't support type [Attention], Attention/Reshape_116_output_0
Don't support type [Attention], Attention/Reshape_116_output_0
Don't support type [Attention], Attention/Reshape_125_output_0
Don't support type [Attention], Attention/Reshape_125_output_0
Don't support type [Attention], Attention/Reshape_134_output_0
Don't support type [Attention], Attention/Reshape_134_output_0
Don't support type [Attention], Attention/Reshape_143_output_0
Don't support type [Attention], Attention/Reshape_143_output_0
Don't support type [Attention], Attention/Reshape_152_output_0
Don't support type [Attention], Attention/Reshape_152_output_0
Don't support type [Attention], Attention/Reshape_161_output_0
Don't support type [Attention], Attention/Reshape_161_output_0
Don't support type [Attention], Attention/Reshape_170_output_0
Don't support type [Attention], Attention/Reshape_170_output_0
Don't support type [Attention], Attention/Reshape_179_output_0
Don't support type [Attention], Attention/Reshape_179_output_0
Don't support type [Attention], Attention/Reshape_188_output_0
Don't support type [Attention], Attention/Reshape_188_output_0
Don't support type [Attention], Attention/Reshape_197_output_0
Don't support type [Attention], Attention/Reshape_197_output_0
Don't support type [Attention], Attention/Reshape_206_output_0
Don't support type [Attention], Attention/Reshape_206_output_0
Don't support type [Attention], Attention/Reshape_215_output_0
Don't support type [Attention], Attention/Reshape_215_output_0
Don't support type [Attention], Attention/Reshape_224_output_0
Don't support type [Attention], Attention/Reshape_224_output_0
Don't support type [Attention], Attention/Reshape_233_output_0
Don't support type [Attention], Attention/Reshape_233_output_0
Don't support type [Attention], Attention/Reshape_242_output_0
Don't support type [Attention], Attention/Reshape_242_output_0
Don't support type [Attention], Attention/Reshape_251_output_0
Don't support type [Attention], Attention/Reshape_251_output_0
Load Module Done!
Clone Decode Module Done!