Output heads on the How to Make AI Models track. The head is the last layer that matches your label: one logit per class, one number for regression, or a token vocabulary for generation. A wrong head makes a correct backbone useless.
This lesson assumes you already worked through Tokenizers.
The idea in practice
Set the head size from the label map. Save the label map in the artifact.
A concrete check
goal = {
'track': 'How to Make AI Models',
'lesson': 'Output heads',
}
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 2-class head for a 5-class label that was one-hot encoded wrong. 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
Write head output size for a 4-class problem and for a price.
Self-check
- Say Output heads 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.