In this, I couldn't understand what shuffling batch epochs meant. I did undergo a formal ML course, so I am familiar with the basic terminology. But I don't understand this batch epoch. Could you help me with that?
Consider you have 10 examples (numbered 1-10), and you make 5 batches of them. A batch is a subset of the total data that is used to train the model at each iteration. When training a model, it's often not feasible to use the entire dataset at once because of memory limitations. Instead, the data is divided into small batches, and the model is trained on each batch sequentially. Essentially, you would iterate over these batches and train the model.
In epoch 1, say your batches were (1,4), (2,3), (5,7), (8,9) and (6,10). Shuffling batches in each epoch would mean that in the next epoch, you will alter the batches as well.
So in epoch 2, your batches may look as follows: (3,6), (4, 10), (1,7), (2,8) and (5,9).
Hope that's what you were looking to understand, and it's clear now :)
Hey,
In this, I couldn't understand what shuffling batch epochs meant. I did undergo a formal ML course, so I am familiar with the basic terminology. But I don't understand this batch epoch. Could you help me with that?
Thank you.
That's a valid question.
Consider you have 10 examples (numbered 1-10), and you make 5 batches of them. A batch is a subset of the total data that is used to train the model at each iteration. When training a model, it's often not feasible to use the entire dataset at once because of memory limitations. Instead, the data is divided into small batches, and the model is trained on each batch sequentially. Essentially, you would iterate over these batches and train the model.
In epoch 1, say your batches were (1,4), (2,3), (5,7), (8,9) and (6,10). Shuffling batches in each epoch would mean that in the next epoch, you will alter the batches as well.
So in epoch 2, your batches may look as follows: (3,6), (4, 10), (1,7), (2,8) and (5,9).
Hope that's what you were looking to understand, and it's clear now :)