how to train on multi-GPUs

[Royzon]

I want to train with multiple GPUs, where should I change?

[ruiming46zrm]

@Roython Use dataParalle , put codes in Learner.py ---> train() like this:

def train(self,conf,epochs):
        self.model.train()       
        conf.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
        self.model = nn.DataParallel(self.model,device_ids=[0,1,2,3])
        self.model.to(conf.device)
        running_loss = 0. 

[Royzon]

Thank you very much, it works.