Batches on the AI Learning track. A batch is the set of examples used for one update. Small batches are noisy. Large batches are smoother and need more memory. The learning rate often needs to change when the batch size changes.
This lesson assumes you already worked through Epochs and passes over the data.
The idea in practice
Pick a batch that fits in memory with room for activations. If you double the batch, do not assume the same learning rate still works.
A concrete check
goal = {
'track': 'AI Learning',
'lesson': 'Batches',
}
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 of 1 on a noisy label makes the updates thrash. 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
Explain one reason a larger batch can change the result besides speed.
Self-check
- Say Batches 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.