Load Model

model.save()

# model's state_dict
print("Model's state_dict:")
for param_tensor in model.state_dict():
	print(param_tensor, "\t", model.state_dict([param_tensor].size())

torch.save(model.state_dict(), os.path.join(MODEL_PATH, "model.pt"))

# 같은 모델의 형태에서 파라메터만 load 
new_model = TheModelClass() new_model.load_state_dict(torch.load(os.path.join(MODEL_PATH, "model.pt")))

#모델의 architecture와 함께 저장
torch.save(model, os.path.join(MODEL_PATH, "model.pt"))
model = torch.load(os.path.join(MODEL_PATH, "model.pt"))

checkpoints

import torch
torch.save({ 'epoch': e,
	'model_state_dict': model.state_dict(),
	'optimizer_state_dict': optimizer.state_dict(),
    	'loss': epoch_loss},
	f"saved/checkpoint_model_... .pt")
    
checkpoint = torch.load(PATH)
model.load_state_dict(checkpoint['model_state_dict'])
optimizer.load_state_dict(checkpoint['optimizer_state_dict'])
epoch = checkpoint['epoch']
loss = checkpoint['loss']

Transfer Learning

Freezing

vgg = models.vgg16(pretrained=True).to(device)
class MyNewNet(nn.Module):
    def __init__(self):
	super(MyNewNet, self).__init__()
    	self.vgg19 = models.vgg19(pretrained=True)
        self.linear_layers = nn.Linear(1000, 1)
	# Defining the forward pass
    def forward(self, x):
	x = self.vgg19(x)
	return self.linear_layers(x)

#Freezing
for param in my_model.parameters():
	param.requires_grad = False
for param in my_model.linear_layers.parameters():
	param.requires_grad = True

Montitoring Tools

Tensorboard

Weights & Biases

Multi-GPU

Hyperparameter Tuning

PyTorch Troubleshooting