You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository was archived by the owner on Jan 24, 2024. It is now read-only.
我已经在gpu上训练完了基于palm的模型,预测的话可以实现请问可以实现在cpu和gpu上都能运行吗?
我把palm依赖于paddle关于cuda的代码注释掉是可以在cpu上跑的。但是要实现cpu和gpu同时运行(留一个use_gpu=True/False)的属性,是需要对paddle.fluid.core_avx'的 'get_cuda_device_count进行修改吗