Models with more than one input type on the How to Make AI Models track. Multimodal models combine text, image, or audio. Each branch needs its own encoder. Fusion can be as simple as concatenating vectors. Late fusion is easier to debug.
This lesson assumes you already worked through Ranking.
The idea in practice
Train a text-only and an image-only baseline. Add fusion only if it wins.
A concrete check
goal = {
'track': 'How to Make AI Models',
'lesson': 'Models with more than one input type',
}
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 complex fusion with no single-modality baseline. 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
Sketch two encoders and where their vectors join.
Self-check
- Say Models with more than one input type 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.