作者:kafka | 来源:互联网 | 2023-05-31 11:37
I'm foolish. I cannot write the one-key train code.I hate myself.so could u update the code to train
I'm foolish. I cannot write the one-key train code.
I hate myself.
so could u update the code to train from scratch ?
该提问来源于开源项目:lukemelas/EfficientNet-PyTorch
What does it mean "trained with one key"? Has anyone successfully reproduced B4 or larger results on GPUs? I mean from scratch. If so, what are the h-params and number of epochs?