Technology sharing

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

Apud PyTorch, exemplar servare et onerare potes, et formare his gradibus sequendo:

1. Salvum exemplum

   Solent duo modi servare exemplar:

   • Serva totum exemplar (including structuram retialem, pondera, etc.);

     torch.save(model, 'model.pth')
     

   • Solus status_dict exemplaris (tantum pondus parametri continens) salvatur. Haec methodus commendatur quia spatium repositionis servat et flexibilior est cum oneratione:

     torch.save(model.state_dict(), 'model_weights.pth')
     

2. exemplar onus

   Correspondentes dupliciter potest onerare exemplar;

   • Si totum exemplar servasti prius, hoc modo directe onerare potes:

     model = torch.load('model.pth')
     

   • Si modo state_dict prius servata est, exemplar necesse est instantia eadem structura qua exemplar originale, et deinde transire.load_state_dict()Modum onerare ponderibus;

     # 实例化一个与原模型结构相同的模型
     model = YourModelClass()
      
     # 加载保存的state_dict
     model.load_state_dict(torch.load('model_weights.pth'))
      
     # 确保将模型转移到正确的设备上（例如GPU或CPU）
     device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
     model.to(device)
     

3. continue disciplina

   Exemplar levato, exercere pergere potes. Fac deminutionem munus et optimizer et statuas eorum recte oneratas (si antea servaveris).Deinde, modo sequere processum consuetum disciplinae

   # 定义损失函数和优化器
   criterion = nn.CrossEntropyLoss()
   optimizer = optim.SGD(model.parameters(), lr=0.01, momentum=0.9)
    
   # 如果之前保存了优化器状态，也可以加载
   optimizer.load_state_dict(torch.load('optimizer.pth'))
    
   # 开始训练
   for epoch in range(num_epochs):
       for inputs, labels in dataloader:
           inputs, labels = inputs.to(device), labels.to(device)
           
           optimizer.zero_grad()
           outputs = model(inputs)
           loss = criterion(outputs, labels)
           loss.backward()
           optimizer.step()
   

Hoc modo exemplar instituere potes unde tandem illud servasti.