examples: Add details and improve descriptions
Thanks for making an example! 🙏Would like to propose some additions if you don´t mind. I can´t comment directly due to the notebook nature, so hopefully this can help. If you are time sensitive, and agree on the proposed additions/changes, happy to submit a contrib.🎺
- Add # Goal/Intro section Cover the full functionality the example is exploring at the beginning of the notebook
Proposal : Example of fine-tuning sentiment analysis classifier based on imbd and distilbert pretrained model , experiment tracking and results metrics analysis with dvclive and dvc api
-
Add section # Initialize git and dvc in 2nd cell why? It gives context about the necessary requirements beyond the libraries install that a data scientist needs to run some other HF examples
-
Change # Dataset for # Dataset and Tokenization in 3rd cell The section includes a Tokenization process
-
Add # Evaluation metrics section For consistency / coherence with respect to the process
-
Possible discussion L28 Describe in comment what
log_modeldoes for user? which seems to be the improvement of the PR? -
Explain # Comparing section Context : unclear right now whats really going on with naming . I´ve lost a few chapters of this exploring other things . Thinking about sharing in #9709 . When exploring this, a question came into my mind, what is the difference in between dvc.api and dvclive ?
Originally posted by @SoyGema in https://github.com/iterative/dvclive/pull/649#pullrequestreview-1567359143