I was able to get 95% accuracy on Banglalekha dataset to recognize Bengali characters using resnet34.
The state of the art is 99%.
Hi I am doing the exercise from the book, that requires the dataset using cats and dogs. It said it should take a few minutes but so far it has taken me over 40 minutes! Are these examples no longer used? Below is a screenshot and a link to the exercise.
01_intro.ipynb - Colaboratory (google.com)
I’d guess you’re not using the GPU runtime.