Hello. I’m getting this error, please help. Thanks
Can you show a convolutional_model code ?
Please don’t publish your code. That is against the rules. The way to debug this is to look at the given “expected values” and find where the Conv2D layer is with 392 trainable parameters. That’s the one that your code has wrong. In your case, you have 6152 trainable parameters, but the same output shape. That most likely means you’ve got the filter size (quite a bit) wrong. That’s filter size as in “kernel size” not the number of filters. The number of filters must be correct otherwise the output shapes would not match, but the kernel size can be wrong and still give the correct shape because it’s “same” padding, right?
Actually you can see which layer it is in the output you show: it’s the very first conv layer after the input. So you know where to look for the issue.
I will check the rules