BEANiE: Beans Expert Artificial Intelligence

Model Description

BEANiE (Beans Expert Artificial Intelligence) is an advanced AI-powered system designed to be the ultimate coffee companion for enthusiasts, baristas, and curious consumers alike. By leveraging the power of the Gemma2 2B model and fine-tuning it with comprehensive coffee-related datasets, BEANiE aims to provide expert knowledge, personalized recommendations, and interactive guidance on all aspects of coffee.

Usage

import transformers
from huggingface_hub import notebook_login
notebook_login()
BASE_MODEL = "williamjeong2/BEANiE"
model = AutoModelForCausalLM.from_pretrained(BASE_MODEL, device_map={"":0})
tokenizer = AutoTokenizer.from_pretrained(BASE_MODEL)

Example

question = "Question"
recommendation = get_game_recommendation(question)
print(f"Question: {question}")
print(f"Response: {recommendation}")
# Q1
Question: ์ปคํ”ผ๋ฅผ ์ตœ์ƒ์˜ ์‹ ์„ ๋„๋กœ ์žฅ๊ธฐ๊ฐ„ ๋ณด๊ด€ํ•˜๊ธฐ ์œ„ํ•ด ์–ด๋–ค ์กฐ๊ฑด์„ ์ฃผ๋Š” ๊ฒƒ์ด ์ค‘์š”ํ• ๊นŒ์š”?
Response: ์ปคํ”ผ๋Š” ์ˆ˜๋ถ„, ์—ด, ๋น›์œผ๋กœ๋ถ€ํ„ฐ ๋ฉ€๋ฆฌ ๋ณด๊ด€ํ•ด์•ผ ์‹ ์„ ํ•จ๊ณผ ํ’๋ฏธ๋ฅผ ์œ ์ง€ํ•ฉ๋‹ˆ๋‹ค. ๋˜ํ•œ ๊ฐ•ํ•œ ๋ƒ„์ƒˆ๋ฅผ ํก์ˆ˜ํ•˜๋Š” ๊ฒฝํ–ฅ์ด ์žˆ์œผ๋ฏ€๋กœ ๋‹ค๋ฅธ ์Œ์‹๋“ค๋กœ๋ถ€ํ„ฐ๋„ ๋–จ์–ด์ ธ ์žˆ์–ด์•ผ ํ•ฉ๋‹ˆ๋‹ค. ๋ƒ‰์žฅ๊ณ  ๋ณด๊ด€์€ ํ”ผํ•˜๋Š” ๊ฒƒ์ด ์ข‹์œผ๋ฉฐ, ๋ฒฝ์ด๋‚˜ ์˜ค๋ธ ๊ทผ์ฒ˜์— ๋ณด๊ด€ํ•˜์ง€ ์•Š๋„๋ก ์ฃผ์˜ํ•ด์•ผ ํ•ฉ๋‹ˆ๋‹ค.

# Q2
Question: ์นดํŽ˜์ธ์€ ์–ด๋–ค ๊ธฐ์ „์œผ๋กœ ํ”ผ๋กœ๋ฅผ ์™„ํ™”์‹œํ‚ค๋Š” ํšจ๊ณผ๊ฐ€ ์žˆ์„๊นŒ์š”?
Response: ์นดํŽ˜์ธ์€ ์ค‘์ถ”์‹ ๊ฒฝ์„ ์ž๊ทนํ•˜์—ฌ ์ž ์„ ์ซ“๊ณ  ๋‡Œ ์†์˜ ๋„ํŒŒ๋ฏผ ๋†๋„๋ฅผ ์ฆ๊ฐ€์‹œ์ผœ ํ”ผ๋กœ๊ฐ์„ ์™„ํ™”ํ•˜๋Š” ๊ฒƒ์œผ๋กœ ์•Œ๋ ค์ ธ ์žˆ์Šต๋‹ˆ๋‹ค.  ๋˜ํ•œ, ์นดํŽ˜์ธ์€ ์•„๋ฐ๋…ธ์‹ ๊ณผ ๊ตฌ์กฐ๊ฐ€ ๋น„์Šทํ•˜์—ฌ ์•„๋ฐ๋…ธ์‹  ์ˆ˜์šฉ์ฒด์— ๊ฒฐํ•ฉํ•˜์—ฌ ์•„๋ฐ๋…ธ์‹ ์˜ ์ž‘์šฉ์„ ์ฐจ๋‹จํ•˜๊ณ , ์ด๋Š” ์ž ์„ ์œ ๋ฐœํ•˜๋Š” ๋ฉ”์ปค๋‹ˆ์ฆ˜์„ ์ €ํ•ดํ•˜์—ฌ ํ™œ๋ ฅ์ด ๋Š๊ปด์ง€๋Š” ํšจ๊ณผ๋ฅผ ๊ฐ€์ ธ์˜ต๋‹ˆ๋‹ค.

# Q3
Question: ์ปคํ”ผ๋ฅผ ํ†ตํ•ด ์นดํŽ˜์ธ์„ ์„ญ์ทจํ•˜๋Š” ๊ฒƒ๋งŒ์œผ๋กœ๋Š” ๋‹ค์ด์–ดํŠธ ํšจ๊ณผ๊ฐ€ ์žˆ๋‹ค๊ณ  ๋‹จ์ •ํ•  ์ˆ˜ ์—†๋‹ค๋Š” ๊ฒƒ์€ ๋ฌด์Šจ ์˜๋ฏธ์ผ๊นŒ์š”?
Response: ์ปคํ”ผ์— ํ•จ์œ ๋œ ์นดํŽ˜์ธ์€ ์‹ ์ง„๋Œ€์‚ฌ๋ฅผ ํ™œ๋ฐœํ•˜๊ฒŒ ํ•˜์—ฌ ์ผ์‹œ์ ์œผ๋กœ ์—๋„ˆ์ง€ ์†Œ๋น„๋Ÿ‰์„ ์ฆ๊ฐ€์‹œํ‚ค๋Š” ํšจ๊ณผ๊ฐ€ ์žˆ๊ธฐ ๋•Œ๋ฌธ์— ๋‹ค์ด์–ดํŠธ์— ๋„์›€์ด ๋œ๋‹ค๊ณ  ์ƒ๊ฐํ•˜๋Š” ์‚ฌ๋žŒ๋“ค์ด ์žˆ์Šต๋‹ˆ๋‹ค. ๊ทธ๋Ÿฌ๋‚˜ ์ปคํ”ผ ์ž์ฒด์˜ ์—ด๋Ÿ‰์ด ๋†’๊ณ , ๊ณผ๋„ํ•œ ์„ญ์ทจ ์‹œ ์˜คํžˆ๋ ค ์‹์š•์„ ์ฆ๊ฐ€์‹œ์ผœ ๋ฐ˜๋Œ€๋กœ ์ฒด์ค‘ ์ฆ๊ฐ€๋กœ ์ด์–ด์งˆ ์ˆ˜ ์žˆ๋‹ค๋Š” ์—ฐ๊ตฌ ๊ฒฐ๊ณผ๋„ ์žˆ๊ธฐ ๋•Œ๋ฌธ์— ๋‹จ์ˆœํžˆ ์นดํŽ˜์ธ ์„ญ์ทจ๋งŒ์œผ๋กœ ๋‹ค์ด์–ดํŠธ ํšจ๊ณผ๋ฅผ ๊ธฐ๋Œ€ํ•˜๊ธฐ ์–ด๋ ต๋‹ค๋Š” ๊ฒƒ์ž…๋‹ˆ๋‹ค.

Training data

  • ko-coffee-QA์„ ์‚ฌ์šฉํ•˜์˜€์Šต๋‹ˆ๋‹ค.
  • ๋ณธ ๋ฐ์ดํ„ฐ๋Š” wikipedia๋ฅผ Gemma 2 27B๋ฅผ ์‚ฌ์šฉํ•˜์—ฌ ๋งŒ๋“  QA ๋ฐ์ดํ„ฐ์…‹์ž…๋‹ˆ๋‹ค.
Downloads last month
5
Safetensors
Model size
2.61B params
Tensor type
FP16
ยท
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.

Model tree for williamjeong2/BEANiE

Base model

google/gemma-2-2b
Finetuned
(169)
this model

Dataset used to train williamjeong2/BEANiE