ztxz16 / fastllm

纯c++的全平台llm加速库,支持python调用,chatglm-6B级模型单卡可达10000+token / s,支持glm, llama, moss基座,手机端流畅运行
Apache License 2.0
3.28k stars 332 forks source link

Windows下Pyfastllm不能正常import #394

Closed helloimcx closed 8 months ago

helloimcx commented 8 months ago

image

环境

windows 10 python 3.10.13 cuda 11.8

helloimcx commented 8 months ago

faq页面,把cudart cublas复制到pyd同一目录下,已解决