Only raw code with no explanation.

Only a model name without metrics.

Tables of metrics, plots such as confusion matrices or learning curves, and interpretation of what worked best and why.

Only a conclusion written before experiments are performed.