Centos ????? ? ??? ?? Pytorch? ???? ?? ??? ??? ?????.

1. Pytorch ??

Pytorch? ????? Anaconda ?? PIP? ??? ? ????.

A. ???? ??

1. Anaconda ???? : ?? Anaconda ? ??? ?? Centos ??? ? Anaconda3 ?? ???? ????????. ?? ???? ?? ??? ??????.

2. ?? ?? ??? : ???? ?? pytorch ?? ?? ??? ??? ???????.

    ?? ?? -n pytorch python = 3.8
   ?? ??? ? ?? ??

3. Pytorch ?? : ??? ? pytorch ???? Conda? ???? Pytorch? ??????. GPU ??? ??? ?? CUDA ? CUDNN? ???? ??? ???? ?? Pytorch ??? ??????. ?? ??? Cuda 11.8 ??? ???? Pytorch? ?????.

    ?? ?? Pytorch Torchvision Torchaudio Cudatoolkit = 11.8 -C Pytorch

4. ?? ?? : Python ??? ??? ???? ?? ??? ???? Pytorch? ????? ?????? ???? GPU ???? ??????.

    ?? ??
   ?? (?? .__ ?? __)
   print (torch.cuda.is_available ())

B. PIP ??

1. PIP ?? : ???? PIP? ???? ?? ?? ?? ??????.

    Sudo Yum ?? Python3-Pip

2. Pytorch ?? : PIP? ???? Pytorch? ???? Tsinghua University Mirror ??? ???? ???? ??? ?????.

    Pip Torch Torchvision Torchaudio -f https://pypi.tuna.tsinghua.edu.cn/simple

3. ?? ?? : Anaconda ???? ???? ?? ??? ???? ??? ??????.

    ?? ??
   ?? (?? .__ ?? __)
   print (torch.cuda.is_available ())

2. ? ?? ??

??? ? ??? Pytorch? ???? ??? ???? ??? MNIST ?? ?? ?? ????.

1. ?? ?? ????? :

    ?? ??
   Torch.nn? nn?? ?????
   Torch.optim? ???? ?????
   TorchVision? ?????
   torchvision.transforms? ???? ?????

2. ?? ?? : ??? ??? ?? ?? ?? ???? (CNN)???.

    ??? simplecnn (nn.module) :
       def __init __ (self) :
           Super (simplecnn, self) .__ init __ ()
           self.conv1 = nn.conv2d (1, 32, kernel_size = 3, ?? = 1)
           self.pool = nn.maxpool2d (2, 2)
           self.fc1 = nn.linear (32 * 14 * 14, 10) #??? ??? ??? defforfort (self, x)? ?? ??? ?????.
           x = self.pool (Torch.relu (self.conv1 (x)))
           X = Torch.flatten (x, 1) # flatten x = self.fc1 (x)
           ?? x

3. ??? ?? : MNIST ??? ??? ?????? ???????.

    transform = transforms.compose ([transforms.totensor (), transforms.normanize ((0.1307,), (0.3081,))))
   Train_dataset = torchvision.datasets.mnist (root = './data', train = true, download = true, transform = transform)
   test_dataset = torchvision.datasets.mnist (root = './data', train = false, download = true, transform = transform)
   Train_Loader = torch.utils.data.dataloader (Train_Dataset, batch_size = 64, ?? = true)
   test_loader = torch.utils.data.dataloader (test_dataset, batch_size = 1000, ?? = ??)

4. ??, ?? ?? ? ???? ???????.

    model = simplecnn ()
   ?? = nn.crossentropyloss ()
   Optimizer = Optim.adam (model.parameters (), lr = 0.001) # Adam Optimizer ??

5. ?? ?? :

    ??? = 2
   ??? (Epochs)? Epoch? ?? :
       running_loss = 0.0
       I? ??, ?? ? ??? (Train_Loader, 0) :
           ??, ?? = ???
           Optimizer.zero_grad ()
           ?? = ?? (??)
           ?? = ?? (??, ???)
           loss.backward ()
           Optimizer.step ()
           running_loss = loss.Item ()
           I % 100 == 99 ? ?? :
               print (f '[{epoch 1}, {i 1}] ?? : {running_loss / 100 : .3f}')
               running_loss = 0.0
   ?? ( '?? ??')

6. ?? ?? :

    ??? = 0
   ? = 0
   Torch.no_grad ()? :
       test_loader? ??? :
           ???, labels = data
           ?? = ?? (???)
           _, ?? = ?? .max (outputs.data, 1)
           Total = labels.size (0)
           right = (?? == ??) .sum (). item ()
   
   print (f'accuracy : {100 * ??? / Total}%')

? ??? ?? ??? ??? ?????. ??? ?? ?? ??, ??? ?? ? ??? ????? ??? ? ????. ???? ?? ./data ????? ???????. ? ??? Adam Optimizer? ???? ????? SGD?? ??? ?????. ??? ??? ???? ?? ??? ?? ??? ? ??? ?? ?????.

? ??? Centos? Pytorch?? ? ??? ???? ??? ?? ?????. ??? ??? PHP ??? ????? ?? ?? ??? ?????!