[fix] Fix data of qk_matmul in prefill stage

hahnyuan / LLM-Viewer

Analyze the inference of Large Language Models (LLMs). Analyze aspects like computation, storage, transmission, and hardware roofline model in a user-friendly interface.

MIT License

310 stars 37 forks source link

[fix] Fix data of qk_matmul in prefill stage #5

Closed jschenxiaoyu closed 6 months ago

jschenxiaoyu commented 6 months ago

In qk_matmul，q belongs to load_act, not load_kv_cache

hahnyuan commented 6 months ago

You are right, I have merged this.