Update src/generation.py

#1

Implement generation.answer() with flan-t5-base

  • Load and cache text2text-generation pipeline (GENERATION_MODEL)
  • Build context from ranked chunks, truncating to MAX_CONTEXT_TOKENS
    to stay within flan-t5-base's context window
  • Keep existing RELEVANCE_THRESHOLD fallback guard unchanged
  • Replaces stub echo of top chunk
huggingfaceconsumer changed pull request status to merged

Sign up or log in to comment