RuntimeError: The size of tensor a (4) must match the size of tensor b (3) at non-singleton dimensio

Use transfroms Method process error ：

RuntimeError: The size of tensor a (4) must match the size of tensor b (3) at non-singleton dimension 0

Code block ：

from PIL import Image
#from torchvision import transforms
# First import a picture 
img = Image.open("images/pytorch.png")
print(img)# Run it to see if the picture is read normally 
# It is PIL Type of picture , The picture size is 955×500

from torchvision import transforms
# First of all to see compose class , Is to put different transforms Combined with 

# Small class 
#python in __call__ Usage of 

# transforms Of compose Methods can be called either through built-in functions or through __call__ To call 
# A class object is often followed by a bracket () Form call of 

#ToTensor
# Will a PIL perhaps numpy Convert image into tensor
from torch.utils.tensorboard import SummaryWriter# from tensorboard Introduced SummaryWriter
writer = SummaryWriter("logs")
img = Image.open("images/pytorch.png")
print(img)


trans_totensor = transforms.ToTensor()
img_tensor = trans_totensor(img)
writer.add_image("ToTensor",img_tensor)
# writer.close()


##Normalization normalization 
#  Can normalize a tnesor Image , With his mean or standard deviation 
#  Give a mean and standard deviation dimension 
#  The value of each channel can be normalized 

#  Normalization formula 
# output[channel] = (input[channel] - mean[channel]) / std[channel]`
print(img_tensor[0][0][0])
trans_norm = transforms.Normalize([0.5, 0.5, 0.5], [0.5, 0.5, 0.5])
img_norm = trans_norm(img_tensor)
print(img_norm[0][0][0])
writer.add_image("normalize",img_norm)
writer.close()

After the error is reported, locate to the 40 Line code ：

img_norm = trans_norm(img_tensor)

Cause analysis ：

The input image has a alpha Pixels of the channel , So it's a four channel image instead of a three channel image

Verify the conjecture

resolvent ：

Connect four channels RGBA The image is converted into a three channel RGB Images

img = Image.open("images/pytorch.png").convert('RGB')