Spike: Synthetic data creation

CDCgov / IDWA

Intelligent Data Workflow Automation

Apache License 2.0

1 stars 1 forks source link

Spike: Synthetic data creation #78

Closed knguyenrise8 closed 1 month ago

knguyenrise8 commented 3 months ago

Create a 1 pger/2 pger on how effective fine tuning llms on parsing pdfs. Provide numbers on time to parse a pdfs. create numbers on 1page, 2page, 3page etc, costs/tokens per transaction.

Look into GPT 3.5/4 OLLAMA etc.

bora-skylight commented 1 month ago

Tasks: @knguyenrise8 please edit!

Synthetic data creation
Use synthetic data to generate sample PDF forms
Use teseract for OCR + LLM

bora-skylight commented 1 month ago

@knguyenrise8 Can you please break down this ticket into 3 separate tickets as we mentioned above?