I have to train vgg16 from scratch on a digit dataset as my college project. I require more computational power. Can anyone suggest to me how can I do that other than Google Colab? Currently, it is taking approx 17 hours to run one epoch? Like on cloud or anything? and it should be cost-efficient too as I'm just a student.
3 Answers
I suggest you can give a try in JOVIAN. Even you can go with Kaggle Notebooks. Also, suggest to activate GPU Runtime in Google Colab.
Cocalc is a nice platform for running notebooks in the cloud. Notebooks are offered for pretty much all languages that can be run in a jupyter notebook.