There seems to be some issue in graph for validation loss in the last assignment (cell 34) for module 4. I have got all the marks for this assignment though the graph doesn’t seem to be correct.
Interesting. Thanks for sharing @tanishk007
I’m going to send you a direct message regarding this.
1 Like
After looking at the notebook, providing the solution for this for others:
The instructions for validation epoch function (exercise 5), particularly for accumulated loss (running_val_loss) are to “get the scalar value from the loss tensor and scale it by the batch size”
Tanishk was not scaling the val loss by the batch size.
For reference, take a look at how it is being done in training epoch function (exercise 4).
2 Likes