summary() missing 8 required keyword-only arguments

Hello, I am trying to use summary() on a simple CNN function. However, I get the following TypeError:

TypeError: summary() missing 8 required keyword-only arguments: 'batch_dim', 'branching', 'col_names', 'col_width', 'depth', 'device', 'dtypes', and 'verbose'

I am using:

from torchsummary import summary
model = CNN().to(device)
summary(model, (3, 96, 96))

What is the correct execution?

[wlcosta]

Well, this is confusing. I restarted the kernel once again and it worked just fine:

==========================================================================================
Layer (type:depth-idx)                   Output Shape              Param #
==========================================================================================
├─Conv2d: 1-1                            [-1, 32, 26, 26]          896
├─Conv2d: 1-2                            [-1, 64, 24, 24]          18,496
├─Linear: 1-3                            [-1, 128]                 1,179,776
├─Linear: 1-4                            [-1, 10]                  1,290
==========================================================================================
Total params: 1,200,458
Trainable params: 1,200,458
Non-trainable params: 0
Total mult-adds (M): 12.38
==========================================================================================
Input size (MB): 0.01
Forward/backward pass size (MB): 0.45
Params size (MB): 4.58
Estimated Total Size (MB): 5.04
==========================================================================================

I am not sure what happened since I already restarted the kernel a few times. Thank you for your help! Best regards.

[wlcosta]

Yeah, it should be (3, 28, 28). I changed it here, but the error persists. I’m still updating my environment to Python 3.8.