Batch size and memory on the How to Train Models track. Batch size is how many examples contribute to one update. Larger batches smooth the gradient and use more memory. Very small batches make noisy updates.
This lesson assumes you already worked through Adam and when it hides problems.
The idea in practice
Fit the batch to hardware with room for activations. If you change batch size a lot, retune the learning rate.
A concrete check
goal = {
'track': 'How to Train Models',
'lesson': 'Batch size and memory',
}
checks = [
'input available at decision time',
'score matches the real decision',
'failure case written down',
]
print(goal['lesson'])
for item in checks:
print('-', item)
Run the sketch locally if you have Python. The printout is a reminder of the checks, not a trained model. Replace the strings with the real inputs from your own example before you treat it as a design.
What usually goes wrong
A batch that barely fits and then crashes when you add a metric. When this happens, stop adding parameters or tools. Fix the check, the data, or the permission, then run the same example again.
What to write down
- The input you are allowed to use at decision time.
- The output and the score or pass rule.
- One failure you will test on purpose.
- What you will not claim the system can do.
Practice
State what you would reduce first if you run out of GPU memory.
Self-check
- Say Batch size and memory in one sentence that mentions an input and an output.
- Name the failure mode in this lesson and the check that would catch it.
Done when: you can explain this lesson without the page open, and you have a written failure case.