I am currently on lesson 8 where Jeremy has explained about the forward and the backward pass. However replicating his notebook I still get out of memory error on mnist data for forward_and_backward pass functions. I haven’t broken my data into batches just like in his notebook. However I am running it on google colab with 25 GB RAM. Can someone please help out?
Can you share your notebook here? It’ll be easier to help you out that way. Also, you could simply try breaking down your data into batches until you don’t get OOM errors? That should work.
Is there a specific reason for doing a full batch training here?
No specific reason, just replicating it as in the lesson, there I don’t see any mini batches