Fine-tune a small open model

Convert train and validation rows to chat format: the email in, the label out

Run this to check

You pass when

One JSON line with a messages list: a user message and a one-word assistant message.

Files you write

• data/train.chat.jsonl
• data/valid.chat.jsonl

Score the untuned small model on the test rows

Run this to check

You pass when

An accuracy number, usually below your classic baseline.

Files you write

• eval_tuned.py

Train an adapter and log training and validation loss every 50 steps

An adapter trains a few million numbers and leaves the model untouched, which is why it fits on a laptop.

Run this to check

You pass when

Both losses went down from the first line of the log. The adapter is a small file, not a new model.

Files you write

• train.log
• adapters/

Find the step where validation loss stops falling and keep that checkpoint

When training loss keeps falling and validation loss turns up, the model is memorising. Stop there.

Run this to check

You pass when

A tuned accuracy on the same 30 rows. In RESULTS.md, say which step you kept and why.

Files you write

• RESULTS.md