styling_gan_model.ipynb Object type return error
insung3511 opened this issue · comments
Error Code
def runStyleTransfer(cnn, contentImg, styleImg, num_steps=300, style_weight=100000, content_weight=1):
inputImg = contentImg.clone().detach().requires_grad_(True)
model, styleLosses, contentLosses = getStyleModelAndLosses(cnn, styleImg, contentImg)
optimizer = optim.LBFGS([inputImg])
iteration = [0]
while iteration[0] <= num_steps:
def closuer():
inputImg.data.clamp_(0, 1)
optimizer.zero_grad()
model(inputImg)
styleScore = 0
contentScore = 0
for sl in styleLosses:
styleScore += sl.loss
for cl in contentLosses:
contentScore += cl.loss
loss = (style_weight * styleScore) + (content_weight * contentScore)
loss.backward()
iteration[0] += 1
if iteration[0] % 50 == 0:
print('Iteration {}: Style Loss {:4f}\tContent Loss : {:4f}'.format((
iteration[0], styleScore.item(), contentScore.item()
)))
return torch.Tensor(styleScore + contentScore)
optimizer.step(closuer)
return inputImg.data.clamp_(0, 1)Error
Output exceeds the [size limit](command:workbench.action.openSettings?[). Open the full output data [in a text editor](command:workbench.action.openLargeOutput?859d898f-236e-4f66-9f8a-135b525cfe5d)
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
Cell In[80], line 1
----> 1 output = runStyleTransfer(cnn, contentImg=contentImg, styleImg=styleImg)
Cell In[78], line 30, in runStyleTransfer(cnn, contentImg, styleImg, num_steps, style_weight, content_weight)
28 return torch.Tensor(styleScore + contentScore)
29 print(type(closuer))
---> 30 optimizer.step(closuer)
31 return inputImg.data.clamp_(0, 1)
File ~/miniconda3/lib/python3.10/site-packages/torch/optim/optimizer.py:140, in Optimizer._hook_for_profile.<locals>.profile_hook_step.<locals>.wrapper(*args, **kwargs)
138 profile_name = "Optimizer.step#{}.step".format(obj.__class__.__name__)
139 with torch.autograd.profiler.record_function(profile_name):
--> 140 out = func(*args, **kwargs)
141 obj._optimizer_step_code()
142 return out
File ~/miniconda3/lib/python3.10/site-packages/torch/autograd/grad_mode.py:27, in _DecoratorContextManager.__call__.<locals>.decorate_context(*args, **kwargs)
24 @functools.wraps(func)
25 def decorate_context(*args, **kwargs):
26 with self.clone():
---> 27 return func(*args, **kwargs)
File ~/miniconda3/lib/python3.10/site-packages/torch/optim/lbfgs.py:438, in LBFGS.step(self, closure)
...
---> 26 iteration[0], styleScore.item(), contentScore.item()
27 )))
28 return torch.Tensor(styleScore + contentScore)
AttributeError: 'int' object has no attribute 'item'
This error happend on optimizer.step(clouser) function, it's working before 50th loop. Still try to solve this problems. If anyone knows about this solutions please comment on this issue. Thanks :)
Still can't get it how to possible object function get a return. Also, I got a really don't know where has got a interger variable in clouser object.
I checked variable type and values.
returnValue = styleScore + contentScore
print(returnValue)
return returnValueand this is result
tensor(7.9810e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(7.9804e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(5.8912e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.0803e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(2.8521e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(3.7512e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.9474e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(3.6544e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(0.0001, device='mps:0', grad_fn=<AddBackward0>)
tensor(3.5965e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.6215e-05, device='mps:0', grad_fn=<AddBackward0>)
tensor(3.3496e-06, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.5352e-06, device='mps:0', grad_fn=<AddBackward0>)
tensor(7.3660e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.1677e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(2.9011e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(2.0193e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.6363e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.4279e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.2995e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.0171e-07, device='mps:0', grad_fn=<AddBackward0>)
tensor(7.7450e-08, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.7257e-08, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.9968e-08, device='mps:0', grad_fn=<AddBackward0>)
tensor(9.0057e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(6.4314e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(5.8167e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(5.8167e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(5.1039e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(5.1039e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.1998e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.1998e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(2.6534e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(4.5921e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.8199e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.7101e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.7101e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.6390e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.6390e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.5769e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.5769e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.4023e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.4023e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.3590e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.3590e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.2115e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.2115e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.1137e-09, device='mps:0', grad_fn=<AddBackward0>)
tensor(1.1137e-09, device='mps:0', grad_fn=<AddBackward0>)
Output exceeds the [size limit](command:workbench.action.openSettings?[). Open the full output data [in a text editor](command:workbench.action.openLargeOutput?a0dfa821-c439-4662-85dd-f9a3ecf32e9f)
---------------------------------------------------------------------------
AttributeError Traceback (most recent call last)
Cell In[17], line 1
----> 1 output = runStyleTransfer(cnn, contentImg=contentImg, styleImg=styleImg)
Cell In[15], line 32, in runStyleTransfer(cnn, contentImg, styleImg, num_steps, style_weight, content_weight)
30 print(returnValue)
31 return returnValue
---> 32 optimizer.step(closuer)
33 return inputImg.data.clamp_(0, 1)
File ~/miniconda3/lib/python3.10/site-packages/torch/optim/optimizer.py:140, in Optimizer._hook_for_profile.<locals>.profile_hook_step.<locals>.wrapper(*args, **kwargs)
138 profile_name = "Optimizer.step#{}.step".format(obj.__class__.__name__)
139 with torch.autograd.profiler.record_function(profile_name):
--> 140 out = func(*args, **kwargs)
141 obj._optimizer_step_code()
142 return out
File ~/miniconda3/lib/python3.10/site-packages/torch/autograd/grad_mode.py:27, in _DecoratorContextManager.__call__.<locals>.decorate_context(*args, **kwargs)
24 @functools.wraps(func)
25 def decorate_context(*args, **kwargs):
26 with self.clone():
---> 27 return func(*args, **kwargs)
File ~/miniconda3/lib/python3.10/site-packages/torch/optim/lbfgs.py:438, in LBFGS.step(self, closure)
...
27 )))
29 returnValue = styleScore + contentScore
30 print(returnValue)
AttributeError: 'int' object has no attribute 'item'
Solved. It wasn't not a big issue with it. While training model, contentLoss become variable int type. Happened this problem at printing iteration, styleScore, contentScore.
Original code was like this.
if iteration[0] % 50 == 0:
print('Iteration {}: Style Loss {:4f}\tContent Loss : {:4f}'.format(
iteration[0], styleScore.item(), contentScore.item()
))and this is changed
if iteration[0] % 50 == 0:
print('Iteration {}: Style Loss {:4f}\tContent Loss : {:4f}'.format(
iteration[0], styleScore.item(), contentScore
))@holopin-bot @insung3511 Can you do some for me?
Looks like @insung3511 is trying to award a sticker, but something went wrong while doing so. See this page for more information.