PGOS Emotion Agent โ€” Stage 1 (๋Œ€๋ถ„๋ฅ˜)

Psychological Growth OS(PGOS) ํ”„๋กœ์ ํŠธ์˜ ๊ฐ์ • ๋ถ„๋ฅ˜ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค. ๊ณ„์ธต์  2๋‹จ๊ณ„ ๋ถ„๋ฅ˜๊ธฐ์˜ 1๋‹จ๊ณ„ ๋Œ€๋ถ„๋ฅ˜ ๋ชจ๋ธ์ž…๋‹ˆ๋‹ค.

๋ชจ๋ธ ๊ฐœ์š”

์ž…๋ ฅ ํ…์ŠคํŠธ๋ฅผ 6๊ฐœ ๊ฐ์ • ๋Œ€๋ถ„๋ฅ˜๋กœ ๋ถ„๋ฅ˜ํ•ฉ๋‹ˆ๋‹ค.

ํด๋ž˜์Šค ํฌํ•จ ๊ฐ์ •
๋ถ„๋…ธ๊ณ„ ๋ถ„๋…ธ, ์ขŒ์ ˆ, ์งœ์ฆ
์Šฌํ””๊ณ„ ์Šฌํ””, ์šฐ์šธ, ํ›„ํšŒ
๋ถˆ์•ˆ๊ณ„ ๋ถˆ์•ˆ, ๊ฑฑ์ •
์ƒ์ฒ˜๊ณ„ ์ƒ์ฒ˜, ์–ต์šธ, ์งˆํˆฌ
๋‹นํ™ฉ๊ณ„ ์ˆ˜์น˜์‹ฌ, ์ฃ„์ฑ…๊ฐ, ๋‹นํ™ฉ
๊ธฐ์จ๊ณ„ ๊ธฐ์จ, ์•ˆ๋„

์„ฑ๋Šฅ

์ง€ํ‘œ ๊ฐ’
Macro F1 0.692
Accuracy 0.684

ํ•™์Šต ๋ฐ์ดํ„ฐ

  • ์›์ฒœ: AI Hub 018 ๊ฐ์„ฑ๋Œ€ํ™” ๋ง๋ญ‰์น˜
  • ๋ฐ์ดํ„ฐ: ํด๋ž˜์Šค๋‹น ๊ท ๋“ฑ ์ƒ˜ํ”Œ๋ง + GPT-4o-mini ์ฆ๊ฐ•
  • ์ด ํ•™์Šต ๋ฐ์ดํ„ฐ: 64,000๊ฐœ

๋ฒ ์ด์Šค ๋ชจ๋ธ

beomi/KcELECTRA-base-v2022

์‚ฌ์šฉ ๋ฐฉ๋ฒ•

from transformers import AutoTokenizer, AutoModelForSequenceClassification
import torch

tokenizer = AutoTokenizer.from_pretrained("nuguri01/pgos-emotion-stage1-major")
model = AutoModelForSequenceClassification.from_pretrained("nuguri01/pgos-emotion-stage1-major")

text = "์˜ค๋Š˜ ๋„ˆ๋ฌด ํ™”๊ฐ€ ๋‚˜๊ณ  ์งœ์ฆ์ด ๋ฐ€๋ ค์™€์š”"
inputs = tokenizer(text, return_tensors="pt", max_length=128, truncation=True)
with torch.no_grad():
    logits = model(**inputs).logits
pred = torch.argmax(logits, dim=-1).item()
labels = ["๋ถ„๋…ธ๊ณ„", "์Šฌํ””๊ณ„", "๋ถˆ์•ˆ๊ณ„", "์ƒ์ฒ˜๊ณ„", "๋‹นํ™ฉ๊ณ„", "๊ธฐ์จ๊ณ„"]
print(f"๋Œ€๋ถ„๋ฅ˜: {labels[pred]}")

PGOS ํŒŒ์ดํ”„๋ผ์ธ

์ž…๋ ฅ ํ…์ŠคํŠธ โ†“ Stage 1 (์ด ๋ชจ๋ธ) โ†’ ๋Œ€๋ถ„๋ฅ˜ ์˜ˆ์ธก โ†“ Stage 2 (์†Œ๋ถ„๋ฅ˜ ๋ชจ๋ธ) โ†’ ์„ธ๋ถ€ ๊ฐ์ • ์˜ˆ์ธก โ†“ Top-3 ๊ฐ์ • ๋ฐ˜ํ™˜ (Hit Rate 0.877)

๊ด€๋ จ ๋ชจ๋ธ

ํ”„๋กœ์ ํŠธ

  • ์„œ๋น„์Šค: Psychological Growth OS (PGOS)
  • ์ฒ ํ•™: Carl Rogers ์ธ๊ฐ„์ค‘์‹ฌ์น˜๋ฃŒ + ACT Hexaflex
  • Behavior Agent: nuguri01/pgos-behavior-agent
Downloads last month
17
Safetensors
Model size
0.1B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support