yeonsue / Problem-solving-with-Donut

OCR Model, GPT API, Streamlit

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

๐Ÿฉ Problem-solving-with-Donut

ํ”„๋กœ์ ํŠธ ์†Œ๊ฐœ

  • ์ด๋ฏธ์ง€๋กœ ๋ฌธ์ œ๋ฅผ ์ž…๋ ฅํ–ˆ์„๋•Œ ์ด๋ฏธ์ง€ ์† ๋ฌธ์ œ์˜ ๋‹ต์„ ์–ป์„ ์ˆ˜ ์žˆ๋Š” ์›นํŽ˜์ด์ง€ ์ œ์ž‘
  • Donut๋ชจ๋ธ, openai api, streamlit์„ ์ด์šฉํ•ด ํ”„๋กœ์ ํŠธ ์ง„ํ–‰
  • Donut๋ชจ๋ธ๋กœ ๋ฌธ์ œ ์ด๋ฏธ์ง€๋กœ๋ถ€ํ„ฐ ๋ฌธ์ œ์™€ ์„ ์ง€๋ฅผ parsingํ•˜์—ฌ openai์˜ GPT-3.5 ๋ชจ๋ธ์— ์ž…๋ ฅํ•˜์—ฌ ๋‹ต๋ณ€ ์ถœ๋ ฅ, streamlit์œผ๋กœ ์›นํŽ˜์ด์ง€ ๊ตฌํ˜„

image

Donut model finetuning

  • dataset : ํ† ์ต ๋ฌธ์ œ 300๊ฐœ, ์ž์ฒด ์ œ์ž‘ํ•œ ๋ฐ์ดํ„ฐ์…‹์œผ๋กœ finetuning ์ง„ํ–‰
  • finetuned model

Getting Started

    pip install .
    pip install timm==0.5.4
    pip install python-dotenv
    pip install langchain
    pip install openai
    pip install streamlit

'streamlit run Toeic.py' ์ž…๋ ฅํ•ด์„œ ์‹คํ–‰

๊ฒฐ๊ณผ ํ™”๋ฉด

result

About

OCR Model, GPT API, Streamlit


Languages

Language:Python 100.0%