Skip to main content

Week 2

Milestones​

  • Generated synthetic data required for finetuning BART.
  • Validate the quality of generated samples.
  • Finetune and evaluate BART on test cases.

Contributions​

  • Updated issue.
  • Commit: Added training pipeline (from generating data to training BART).
  • Commit: Reported results on finetuned BART model.

Learnings​

  • Learnt more about prompt engineering methods and using Hugging Face tools.
  • Read and understood DSP paper.