during training it seem like no working to gpu?

WeidiXie / VGG-Speaker-Recognition

Utterance-level Aggregation For Speaker Recognition In The Wild

364 stars 97 forks source link

during training it seem like no working to gpu? #4

Closed eeric closed 5 years ago

eeric commented 5 years ago

src/main.py It didn't work to gpu ,did it?

eeric commented 5 years ago

it meaned no gpu to train resnet model.

WeidiXie commented 5 years ago

please check https://github.com/WeidiXie/VGG-Speaker-Recognition/blob/fdd2856f590c17f761c9aab9f9ca13de554d06d4/src/main.py#L41

https://github.com/WeidiXie/VGG-Speaker-Recognition/blob/fdd2856f590c17f761c9aab9f9ca13de554d06d4/tool/toolkits.py#L4

eeric commented 5 years ago

I saw it, but feature array wasn't accelerated by gpu.

WeidiXie commented 5 years ago

I tested the code several times on several machines (Linux), the gpus all work fine, so I would suggest to check the installation of the TF and Keras on your machine.

eeric commented 5 years ago

ok, i only readed train code, not run.