well what is the use of batch size?? one thing i forget to ask in previos video is in image classification we devide our data with 255… what is this 255 and why we devide??
What is batch size?
Usually we have very huge datasets which we pass through our models, these datasets have a huge RAM and memory requirement. Thus, to avoid memory based errors and for faster computation we pass our data through our ML Model in form of batches.
So basically, our entire data is passed in form of batches (a group of 32 or 64 or 128… depending upon the data size). With every epoch the entire data is passed and within every epoch there are (total size of data)/(batch_size) number of steps being taken.
I hope this resolves your doubt.
Have answered this in your other doubt.