In Mini batch video, when calculating the loss pratik bhaiya forgot to use the batch dataset and calculated the loss for the entire dataset inside the mini-batch loop and the curve for that was not noisy? But in theory, it should be noisy as well.
In Mini batch video, when calculating the loss pratik bhaiya forgot to use the batch dataset and calculated the loss for the entire dataset inside the mini-batch loop and the curve for that was not noisy? But in theory, it should be noisy as well.