feat(training): update fine tuning examples to our API, linting v1 (#43)

I also updated prepare_for_training to add a more sophisticated "linting" step (example output shown in the screenshot) as an experiment - my plan is to think about that more in the scope of: https://www.notion.so/kazm/Training-Workflow-pipeline-linter-3288e5065c7580be8629ecabf8e4f217?source=copy_link

![image.png](https://app.graphite.com/user-attachments/assets/c4b8c4c2-eb2a-48a0-828e-c94e9872cd20.png)

Author: bartolomej