Transfer from another task on the AI Learning track. Transfer uses a model trained on a related task as a starting point. It helps when you have little data and the old features are relevant. It hurts when the old task teaches the wrong features.
This lesson assumes you already worked through Reproducibility.
The idea in practice
Freeze early layers first, train the head, then unfreeze if validation improves. Compare against a model trained only on your data.
A concrete check
goal = {
'track': 'AI Learning',
'lesson': 'Transfer from another task',
}
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
Fine-tuning a huge model on fifty rows with a high learning rate and calling it transfer. 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
Say when you would rather train a small model from scratch.
Self-check
- Say Transfer from another task 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.