CUDA多卡调用

调用库上方添加:

os.environ['CUDA_VISIBLE_DEVICES']= '0,1'

设备设定由:

device = torch.device(args.device if torch.cuda.is_available() else "cpu")

改为:

device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')

模型加载设定由:

model = create_model(num_classes=args.num_classes).to(device)

改为:

model = create_model(num_classes=args.num_classes)

外加:

model = nn.DataParallel(model,device_ids=None).to('cuda')

原代码:

device = torch.device(args.device if torch.cuda.is_available() else "cpu")
model = create_model(num_classes=args.num_classes).to(device)

现代码:

os.environ['CUDA_VISIBLE_DEVICES']= '0,1'
device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
model = create_model(num_classes=args.num_classes)
model = nn.DataParallel(model,device_ids=None).to('cuda')

你可能感兴趣的:(人工智能,linux,python,服务器)