Fine-tuned using this notebook

Sample Usage


# Load the model
import torch
from peft import PeftModel, PeftConfig
from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
 
### Load peft config for pre-trained checkpoint etc.
peft_model_id = "results"
config = PeftConfig.from_pretrained(peft_model_id)
 
### Load base LLM model and tokenizer
model = AutoModelForSeq2SeqLM.from_pretrained("divakaivan/t5-large-finetuned-reviewer-kr",  load_in_8bit=True,  device_map={"":0})
tokenizer = AutoTokenizer.from_pretrained("divakaivan/t5-large-finetuned-reviewer-kr")
 
### Load the Lora model
model = PeftModel.from_pretrained(model, peft_model_id, device_map={"":0})
model.eval()

# Sample Input

### Review
review = """๋ฆฌ๋ทฐ 1: ๋ถˆํ›„์˜ ์„ฑ๋Ÿ‰์œผ๋กœ ์ œ๋ชฉ ๋ฐ”๊ฟ”๋ผ. ์ด๊ฑด ๋ญ ์†Œ๋ฆฌ๋งŒ ๊ฝฅ๊ฝฅ ์ง€๋ฅด๊ณ  ์ง€๋“ค๋ผ๋ฆฌ ๊ฐ๋™์ด๋ž˜. ์ „ํ˜€ ๊ณต๊ฐ ๋ชป ํ•˜๊ฒ ๊ตฌ๋งŒ ์ฏง;  ๋ฆฌ๋ทฐ 2: ๋‚œ ์›๋ž˜ ๊ธฐ๋ฆฐ(์ด๊ด‘์ˆ˜) ์‹ซ์–ดํ•œ๋‹ค ํฌํฌํฌ.... ๋Šฅ๋ ฅ์ž๊ฐ€ ์ข‹๋‹ค;  ๋ฆฌ๋ทฐ 3: ๊ทธ๋ž˜๋„ ๊ฒฐ๊ณผ๋ฅผ ๋ณด๋˜ ๋“œ๋ผ๋งŒ๋ฐ์š”...์ด์„ ๋‚จ์ด๋ž‘ ๋ฏผ์ง€์ˆ˜๊ฐ€ ์ž˜๋˜๊ธธ ๋ฐ”๋ž˜๋Š” ๋งˆ์Œ์—์„œ์š”..๊ทผ๋ฐ ์˜ค๋Š˜ ์•„์นจ๋“œ๋ผ๋งŒ ์ •๋ง ์‹ค๋งํ–ˆ์–ด์š”. ๋ˆˆ๋ฌผ์ด ํ•‘๋Œ๋”๋ผ๊ตฌ์š”..์–ด์ฉœ ํ•œํ‰์ƒ ๋‹นํ•˜๊ณ  ์ฐฉํ•˜๊ฒŒ๋งŒ ์‚ฌ๋žŒ์€ ์ฃฝ์–ด์•ผํ•˜๋‚˜์š”.. ์ด๋“œ๋ผ๋งˆ์ž‘๊ฐ€ ์ตœ์•…์˜ ์ตœ์•…์ด์˜ˆ์š”.. ๋“œ๋ผ๋งˆ๋‹ค์‹ ๋ณด๊ณ ์‹ถ์ง€์•Š์•„์š”;  ๋ฆฌ๋ทฐ 4: ๋๊นŒ์ง€ ์•‰์•„์žˆ๊ธฐ๊ฐ€ ๋„˜ ํž˜๋“ค์—ˆ์Œ.. ์žฌ๋ฏธ๋„.. ๋‚ด์šฉ๋„.. ๊ฐ๋™๋„.. ์ตœ์•…์ž„..;  ๋ฆฌ๋ทฐ 5: ๊ทธ๋ƒฅ ์ธ๊ฐ„๊ทน์žฅ์ด ๋” ๋‚ ๋“ฏ..ํ‹ฐ๋น„๋กœ๋ณด๊ตฌ๋งŒ๋‹ค;  ๋ฆฌ๋ทฐ 6: ์‚ฌ๊ณ ๋ญ‰์น˜์ด์ง€๋งŒ ์–ด๋จธ๋‹ˆ์— ๋Œ€ํ•œ ๋œจ๊ฑด์šด ์‚ฌ๋ž‘์ด ๋ˆŒ๋ฌผ๊ฒจ์›Œ ๊ฐ๋™์ ์ด์—ˆ๋‹ค.;  ๋ฆฌ๋ทฐ 7: ์„ค์žฅ์šฐ ์นœ๊ตฌ์—ญ์œผ๋กœ ๋‚˜์˜จ ๋ฐฐ์šฐ๋ถ„ ๋ˆ„๊ตฌ์ง€? ์ œ์ผ ์›ƒ๊ฒผ์Œ;  ๋ฆฌ๋ทฐ 8: ์€ํ˜„์ˆ˜ ๋„ˆ๋ฌด ๋‹ต๋‹ตํ•œ์บ๋ฆญํ„ฐ ๋ณด๊ณ ์žˆ์Œ ์งœ์ฆ๋‹์•„์š”ใ…กใ…ก์•„ ์—ฐ๊ธฐ๋„ ๊ทธ๋‹ฅ;  ๋ฆฌ๋ทฐ 9: ์ด์˜ํ™”๋ณด๊ณ ์‹ถ์€๋ฐ ใ… ใ…  ์–ด๋””์„œ ๋ด์š”? ใ… ใ…  ์ฐพ์„์ˆ˜๊ฐ€์—†์–ด์š”ใ… ใ… ;  ๋ฆฌ๋ทฐ 10: ์ด ์˜ํ™”ํ•œ๋ฒˆ ๋”๋ณผ๋ ค๊ณ ,์ŠคํŒŒ์ด๋”๋งจ,์ฒซ์‚ฌ๋ž‘,์กฐํญ2..๋ณด์ง€๋„ ์•Š์•˜๋‹ค!;  ๋ฆฌ๋ทฐ 11: ์ฝ”์ฝ”๋ชฝ๋‚˜์™”์„๋•Œ๋„ ์šธ์—ˆ์Šต๋‹ˆ๋‹ค..;  ๋ฆฌ๋ทฐ 12: ์•„์ง์•ˆ๋ดค์ง€๋งŒ ๋งˆํ‹ด๋•Œ๋ฌธ์—๋ณธ๋‹คใ…‹ใ…‹ใ…‹ใ…‹;  ๋ฆฌ๋ทฐ 13: ์˜ํ™”์ƒ์˜ ๋‚ด๋‚ด ๊ณณ๊ณณ์—์„œ ํ„ฐ์ ธ๋‚˜์˜ค๋Š” ์‹ค์†Œ. ์ฝ”๋ฏธ๋””๋กœ ์ƒ๊ฐํ•˜๊ณ  ๋ณด๋ฉด ์˜์™ธ๋กœ ๊ดœ์ฐฎ์„์ง€๋„;  ๋ฆฌ๋ทฐ 14: ์ผ์ ๋„ ์•„๊น๋‹ค ์ง„์งœOOO์˜ํ™” ใ…‹ใ…‹ใ…‹ใ…‹ ์•Œ๋ฐ”์•ผ์†”์งํžˆ ์ด๊ฑด์–‘์‹ฌ์—์ฐ”๋ฆฌ์ง€์•Š๋‹ˆ?;  ๋ฆฌ๋ทฐ 15: ์„น์‹œํ•œ๊ฑด ๋‘˜์งธ์น˜๊ณ  ์นด์Šค๋ฏธ๊ฐ€ ์ „ํ˜€ ๋”ดํŒ์ด์ž–์•„ใ…กใ…ก;  ๋ฆฌ๋ทฐ 16: ๋‚˜๋ฅผ ์œ„ํ•œ ์ด์•ผ๊ธฐ .;  ๋ฆฌ๋ทฐ 17: ๋ผ๋ฏผ, ์‹œ์—๋ผ, ํ•˜๋“ค๋ฆฌ ๋‚˜์˜ค๋Š” ์ค„ ์•Œ๊ณ  DVD ๊ตฌ๋งคํ–ˆ๋‹ค๊ฐ€ ๋ˆˆ๋ฌผํ˜๋ฆผ. LND ๋…ธ๋ž˜๋Š” ์ข‹์€๋ฐ ์–˜๋„ค๊ฐ€ ๋ชป๋ถ€๋ฅด๊ณ  ์—ฐ๊ธฐ๋„ ๋ชปํ•จ;;;  ๋ฆฌ๋ทฐ 18: ์ตœ๊ณ ์˜ ์˜ํ™”. ์ด ์˜ํ™”๋ฅผ ๋ณด๊ณ ์„œ ํ˜๋ฆฌ๋Š” ๋ˆˆ๋ฌผ์€ .... ์“ฐ๋ฉด์„œ ๋‹ฌ๋‹ค.;  ๋ฆฌ๋ทฐ 19: ํ†ต์ œ๊ฐ€ ์•ˆ ๋˜๋‹ˆ ์‚ฐ์œผ๋กœ ๊ฐ€๋‹ค.;  ๋ฆฌ๋ทฐ 20: ์žฌ๋ฐŒ๊ฒŒ ๋ดค๋˜ ์ถ”์–ต์˜ ๋“œ๋ผ๋งˆ;  ๋ฆฌ๋ทฐ 21: ๋ป”ํ•œ๋ฐ˜์ „ ,;  ๋ฆฌ๋ทฐ 22: ๋ง์ด ํ•„์š”์—†๋Š” ์—ญ๋Œ€ ์ตœ๊ณ ์˜ ๋ช…์ž‘์ž…๋‹ˆ๋‹ค.;  ๋ฆฌ๋ทฐ 23: ๋ณ„ ํ•œ๊ฐœ๋„ ์•„๊นŒ์šด ์“ฐ๋ ˆ๊ธฐ์˜ํ™”... ์งœ์ฆ๋‚˜;  ๋ฆฌ๋ทฐ 24: ใ… ใ……ใ… )bbbb;  ๋ฆฌ๋ทฐ 25: ์ž์‹ ์„ ๋ฐ”๋กœ๋ณด๊ณ  ์‚ถ์„ ์„ ํƒํ•˜๋Š” ๊ฒƒ์ด ๋‚˜์—๊ฒŒ๋„ ์ฃผ๋ณ€์ธ์—๊ฒŒ๋„ ํ–‰๋ณต์ด ์•„๋‹๊นŒ..์‹ถ์—ˆ๋‹ค.๋‚ด๊ฐ€ ๋‚ด์‚ถ์— ๋งŒ์กฑํ•˜์ง€ ๋ชปํ•˜๊ณ  ๋ถˆํ–‰ํ•˜๋ฉด ๋‚ด์ฃผ์œ„๋„ ๊ฒฐ์ฝ” ํ–‰๋ณตํ•  ์ˆ˜ ์—†๋‹ค.์ด๊ธฐ์ ์ผ๊นŒ..์•„๋‹ˆ ์ˆ™๋ช…์ด๋‹ค.;  ๋ฆฌ๋ทฐ 26: ์ œ๋ชฉ๋ถ€ํ„ฐ ๋ฐ”๊ฟ”๋ผ ํƒ€๊ณ„ํ•˜์‹  ๋ ˆ์ด ๋ธŒ๋ž˜๋“œ๋ฒ„๋ฆฌ๊ฐ€ ๊ทธ๋ ‡๊ฒŒ ์—ญ์ •์„ ๋ƒˆ๋Š”๋ฐ๋„ ๋ป”๋ป”ํ•˜๋„ค;  ๋ฆฌ๋ทฐ 27: ์Šค๋ฆด๋Ÿฌ ํŒฌ์ด์ง€๋งŒ ์ด๊ฑด ๋ญ ์•„๋‹ˆ๋„ค์š”...;  ๋ฆฌ๋ทฐ 28: ์„ฑ์žฅ์˜ํ™”๋ผ๋Š” ํ‹€์— ๊ฐ€๋‘๋ฉด ์ดŒ์Šค๋Ÿฌ์›Œ์ง„๋‹ค. ์ง์„ค์–ด๋ฒ•์„ ํ”ผํ•œ ํ˜„์‹ค๊ณผ ๊ฟˆ, ์•„ํŒŒ์„œ ์ข‹์•˜๋‹ค.;  ๋ฆฌ๋ทฐ 29: ๊ฟ€์žผ ํ—ˆ๋‹ˆ์žผ ใ… ใ…  ๋ณด๋Š”๋‚ด๋‚ด ์›ƒ์Œ์ง€์œผ๋ฉด์„œ๋ณธ ์˜ํ˜ธใ…“ ใ…œใ…œ ๊ฒ์ž ์—ฐ๊ธฐ์ž˜ํ•˜๋„ค;  ๋ฆฌ๋ทฐ 30: ์ด๊ฑฐ ๋ณด๊ณ  ์กฐ์Šน์šฐ๊ฐ€ ๋” ์ข‹์•„์ง..;  ๋ฆฌ๋ทฐ 31: ๊ฒŒ์ž„ ์‹ฌ์ฆˆํ•˜๋Š” ๊ฒƒ ๊ฐ™์ด ์•„๊ธฐ์ž๊ธฐํ•จ. ํฐ ์žฌ๋ฏธ๋Š” ์•„๋‹ˆ์ง€๋งŒ ์ฒ˜์Œ๋ถ€ํ„ฐ ๋๊นŒ์ง€ ์†Œ์†Œํ•˜๊ฒŒ ์œ ์พŒํ•จ.;  ๋ฆฌ๋ทฐ 32: ๋จน๋ฐฉ๋„ ์•„๋‹ˆ๊ณ  ์ณ๋ฌต์ณ๋ฌต ํ•˜๋Š” ์žฅ๋ฉด๋งŒ ๋„ˆ๋ฌด ํด๋กœ์ฆˆ์—… ํ•ด์„œ ์˜ค๋ž˜๋ณด์—ฌ์ฃผ๋‹ˆโ€ฆโ€ฆ. ๋ญ” ๋“œ๋ผ๋งˆ๊ฐ€ ์ด๋Ÿฐ์ง€?;  ๋ฆฌ๋ทฐ 33: 2015๋…„์— ๋ณด๋‹ˆ ๋„ˆ๋ฌด ํ…Œ๋Ÿฌ๋ธ”ํ•˜๋‹ค..;  ๋ฆฌ๋ทฐ 34: ์žฌ๋ฏธ์žˆ๋‹ค. ์—ฌ์„ฑ๊ฐ๋…์œผ๋กœ์„œ ์Šค์ผ€์ผ ํฐ ์ž‘ํ’ˆ์„ ๋งŒ๋“ค ์ˆ˜ ์žˆ๋Š” ๋Šฅ๋ ฅ!;  ๋ฆฌ๋ทฐ 35: ์†Œ์‹ฌํ•œ ๋ณต์ˆ˜, ๊ทธ๋Ÿฌ๋‚˜ ๋„ˆ๋ฌด ํ†ต์พŒ;  ๋ฆฌ๋ทฐ 36: ํ•œ๊ตญ์—๋„ ์ด๋Ÿฐ ์˜ํ™”๊ฐ€ ์žˆ๊ตฌ๋‚˜ .. ๋ณด๊ณ ๋„ ๋ฏฟ์„ ์ˆ˜๊ฐ€ ์—†๋‹คํ•œ ์žฅ๋ฉด ํ•œ ์žฅ๋ฉด ๋‹ค ์ฃผ์˜ฅ๊ฐ™์€ ๋Œ€์‚ฌ๋“ค๊ณผ ๋”๋ถˆ์–ด ๋‚ด ์‚ถ์„ ๋˜๋Œ์•„ ๋ณด๊ฒŒ ๋งŒ๋“œ๋Š” ์˜ํ™”์ธ๊ฒƒ ๊ฐ™๋‹ค;  ๋ฆฌ๋ทฐ 37: ํ•œํฌ์ •์ด๋‹ค...........;  ๋ฆฌ๋ทฐ 38: ์ƒ๊ฐ๋ณด๋‹ค๋„ˆ๋ฌด๊ดœ์ฐฎ์•˜๋˜์˜ํ™”, ์—ฐ๊ธฐํŒŒ ๋ฐฐ์šฐ๋“ค์˜ ์—ฐ๊ธฐ๊ฐ€ ํŠนํžˆ ๋‹๋ณด์ธ๋‹ค;  ๋ฆฌ๋ทฐ 39: ๊ฑฐ์ง“๋ง์Ÿ์ด๋“ค์ด ๋งŒ๋“  ์˜ํ™”;  ๋ฆฌ๋ทฐ 40: ๋นจ๋ฆฌ ์  ์ฐ์–ด....๋นจ๋ฆฌ;  ๋ฆฌ๋ทฐ 41: ํ•œ ๋ฒˆ ์ƒ๊ฐ ํ•ด๋ด. ์†Œ์ˆ˜์˜ ์œ ํƒœ์ธ๋“ค์ด ์–ด๋–ป๊ฒŒ ๋…์ผ ๊ฒฝ์ œ๋ฅผ ๊ฐ‰์•„ ๋จน์—ˆ๋Š” ์ง€๋ฅผโ€ฆ.;  ๋ฆฌ๋ทฐ 42: ๊ฝƒ๋ณด๋‹ค ์†Œ๋…„๋“ค โ™ฅโ™ฅโ™ฅโ™ฅ;  ๋ฆฌ๋ทฐ 43: ๊ธฐ๋ถ„์ด ์ข‹์•„์ง€๋„ค์š”^^;  ๋ฆฌ๋ทฐ 44: ์ผ๋ฐฉ์ , ํŽธํ˜‘ํ•œ ์‚ฌ๊ณ ๋“ค์ด ๋งŒ๋“ค์–ด๋‚ธ ์ผ๋ฐ˜ํ™”์™€ ๊ฐ์ธํšจ๊ณผ๋กœ ์ธํ•ด ํ•œ ์ธ๊ฐ„์˜ ์‚ถ์ด ์–ด๋–ป๊ฒŒ ๋ฐ”๋€”์ˆ˜ ์žˆ๋Š”์ง€ ์•Œ์ˆ˜ ์žˆ๋Š” ์˜ํ™”๋‹ค. ์˜คํžˆ๋ ค ์–ด๋ฆฐ์•„์ด๋‹ˆ๊น ๋˜๋Š” ์‚ฌํšŒ์  ์•ฝ์ž์ด๊ธฐ์— ๋‹น์—ฐํ•˜๊ณ  ๋งž์„๊ฒƒ์ด๋ผ๋Š” ์ƒ๊ฐ์˜ ์˜ค์ ์ด ์–ผ๋งˆ๋‚˜ ํฐ ์‹ค์ˆ˜๊ฐ€ ๋  ์ˆ˜ ์žˆ๋Š”์ง€ ์•Œ์•„์•ผ ํ•  ๊ฒƒ์ด๋‹ค.;  ๋ฆฌ๋ทฐ 45: "์•„๋ จํ•˜๋„ค์š”, ""๊ฑฐ์ ˆ ํ• ํ…Œ๋‹ค."" ""์ด์ œ ํ—ค์–ด์ง€์ž"" ""๋„ค"";  ๋ฆฌ๋ทฐ 46: ๊ฒฌ์ž๋‹จ ์•ก์…˜๋งŒ์œผ๋กœ ์ตœ๊ณ ์˜€์Œ...;  ๋ฆฌ๋ทฐ 47: ์™„์ „ ๋Œ€๋ฐ• ์งฑ ์žผ์žˆ์—ˆ๋‹ค. ๊ทธ๋ฆฌ๊ณ  ์•„.... ๋ฐ˜์ „.... ์ง„์งœ ์žฅ๋‚œ์•„๋‹ˆ๋‹ค.;  ๋ฆฌ๋ทฐ 48: ์ด๊ฑฐ 1์ ์ค€๊ฒƒ๋“ค์€ ๋”๋Ÿฌ์šด OO์ž์‹์ธ๋“ฏ...;  ๋ฆฌ๋ทฐ 49: ์žฅ๊ตญ์˜์ด ์—†๋Š” ์˜์ฑ„์‹  ๋ณ„๋กœ ์˜€๋‹ค. ใ…กใ…ก;  ๋ฆฌ๋ทฐ 50: ใ…† ใ…‚ .... OOO๊ฐ™์€ OOO ์˜ํ™”;"""

### Prompt
prompt = "Your task is to summarise. You are a helpful assistant that helps me evaluate Korean reviews. For each movie you are given 50 reviews. Analyze the reviews, and for the movie itself return a score(1 to 10) and explanation for each of the following criteria: Emotional, Characters, Plot, Visuals, Pacing. Return the review in Korean."

### Model Input
model_input = prompt + review

# Run Inference

input_ids = tokenizer(model_input, return_tensors="pt", truncation=True).input_ids.cuda()
outputs = model.generate(input_ids=input_ids, max_new_tokens=1000, do_sample=True, top_p=0.9)
output = tokenizer.batch_decode(outputs.detach().cpu().numpy(), skip_special_tokens=True)[0]
 
print(output)
Downloads last month

-

Downloads are not tracked for this model. How to track
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 divakaivan/t5-large-finetuned-reviewer-kr

Finetuned
(1)
this model

Dataset used to train divakaivan/t5-large-finetuned-reviewer-kr