how to train model in multiple GPUs

sicxu / Deep3DFaceRecon_pytorch

Accurate 3D Face Reconstruction with Weakly-Supervised Learning: From Single Image to Image Set (CVPRW 2019). A PyTorch implementation.

MIT License

1.7k stars 315 forks source link

how to train model in multiple GPUs #144

Open chenhao-user opened 1 year ago

chenhao-user commented 1 year ago

How to train this in multiple GPUs? I tried using multiple GPUs to train the model via DataParallel , but DataParallel will set self.net_recog.training to True，which causes an error to be reported when running the compute_losses_self。