Skip to content

Add end-to-end Sanskrit poetry verifier #4

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: master
Choose a base branch
from

Conversation

mycpuorg
Copy link

  • Implements dataset generation for Sanskrit poetry prompts with specified meters and topics.
  • Uses gpt-4o to generate verses and a o4-mini for semantic grading to reduce reward hacking.
  • Verifies meter using chandas library with robust Devanagari-to-IAST normalization.
  • Extracts verse lines from LLM output, analyzes meter, syllables, and semantic quality.
  • Saves and analyzes results, including meter-wise performance statistics.
  • Provides a reproducible evaluation pipeline for Sanskrit metrical poetry generation.

Sample output is as shown.

test_output

- Implements dataset generation for Sanskrit poetry prompts with specified meters and topics.
- Uses OpenAI LLM to generate verses and a separate "thinking" model for semantic grading to reduce reward hacking.
- Verifies meter using chandas library with robust Devanagari-to-IAST normalization.
- Extracts verse lines from LLM output, analyzes meter, syllables, and semantic quality.
- Saves and analyzes results, including meter-wise performance statistics.
- Provides a reproducible evaluation pipeline for Sanskrit metrical poetry generation.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant