Assignment: model does not coverge

[Mirza_Canovic]

Hey guys,

I have been trying to finish the assignment for week 4 of this course but I just can’t seem to get the model to converge. I copied the model from the cats vs dogs CAM notebook and replaced the last layer as descriped in the assignment. Also I think that I chose the correct loss function. However, when I train, the loss always goes to ~0.7 and stays there. When I load the pre-trained weights and use them, the loss is larger than 1 at the beginning, then again it quickly goes to ~0.7 and doesn’t change anymore.

Can somebody give me a hint here as to what I could be doing wrong? I have spent the whole day trying to figure out what’s the issue and now I am out of ideas.

Thanks in advance!

[ai_curious]

My training output looks like this…

Did you validate your model summary with the expected?

My loss function is the one that includes this in the documentation…

*Use this cross-entropy loss for binary (0 or 1) classification applications. The loss function requires the following inputs:

• y_true (true label): This is either 0 or 1.*

[Mirza_Canovic]

Alright I don’t know what the issue was but I just ran it again without changing anything and it worked.

Thanks for taking the time though