ValueError: y_pred and y should have same shapes

[oscarramirezs]

Hello

I'm using the "spleen_segmentation_3d.ipynb"on my dataset.

My Shoulder dataset included images and labels (0:background, 1:Humerus , 2: Scapula)

I changed model as
model = UNet(
dimensions=3,
in_channels=1,
out_channels=3,
channels=(16, 32, 64, 128, 256),
strides=(2, 2, 2, 2),
num_res_units=2,
norm=Norm.BATCH,
).to(device)
and
post_pred = Compose([EnsureType(), AsDiscrete(argmax=True, to_onehot=3)])
post_label = Compose([EnsureType(), AsDiscrete(to_onehot=3)])

When I run code, the training occurs well but, it occured a bug in valiadation process: ValueError: y_pred and y should have same shapes

---

C:\ProgramData\Anaconda3\lib\site-packages\monai\metrics\meandice.py in compute_meandice(y_pred, y, include_background)
121
122 if y.shape != y_pred.shape:
--> 123 raise ValueError("y_pred and y should have same shapes.")
124
125 # reducing only spatial dimensions (not batch nor channels)

ValueError: y_pred and y should have same shapes

---

i dont know why the

val_labels.shape is = torch.Size([3, 1, 130, 130, 97])
while the
val_outputs.shape is = torch.Size([1, 3, 130, 130, 97])

can someone help me? I'll appreciate it very much

thanks

[Yangtting]

Please update to run with MONAI v0.8.0 again. I hope I can help you.

[HripsimeS]

@oscarramirezs Hello. Did you find a solution for your issue? I am having the same problem during testing process.

[oscarramirezs]

Hello

I update monai to version 8 and it works fine.