Spaces:
Sleeping
Sleeping
File size: 2,211 Bytes
bda145e |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 |
import gradio as gr
from transformers import AutoTokenizer, AutoModelForSequenceClassification, AutoModelForCausalLM
import torch
import torch.nn.functional as F
# ๊ฐ์ ๋ถ์์ฉ ๋ชจ๋ธ
emotion_model = AutoModelForSequenceClassification.from_pretrained("beomi/KcELECTRA-base", num_labels=3)
emotion_tokenizer = AutoTokenizer.from_pretrained("beomi/KcELECTRA-base")
emotion_labels = ['๋ถ์ ', '์ค๋ฆฝ', '๊ธ์ ']
# ํ
์คํธ ์์ฑ์ฉ GPT ๋ชจ๋ธ
gpt_model = AutoModelForCausalLM.from_pretrained("skt/kogpt2-base-v2")
gpt_tokenizer = AutoTokenizer.from_pretrained("skt/kogpt2-base-v2")
# ๊ฐ์ ๋ถ์ ํจ์
def predict_emotion(text):
inputs = emotion_tokenizer(text, return_tensors="pt", truncation=True, padding=True)
with torch.no_grad():
outputs = emotion_model(**inputs)
probs = F.softmax(outputs.logits, dim=1)
pred = torch.argmax(probs, dim=1).item()
return emotion_labels[pred]
# GPT ์ด์ด์ฐ๊ธฐ ํจ์
def emotional_gpt(user_input):
emotion = predict_emotion(user_input)
if emotion == "๊ธ์ ":
prompt = "๊ธฐ๋ถ ์ข์ ํ๋ฃจ์๋ค. "
elif emotion == "๋ถ์ ":
prompt = "์ฐ์ธํ ๊ธฐ๋ถ์ผ๋ก ์์๋ ํ๋ฃจ, "
else:
prompt = "ํ๋ฒํ ํ๋ฃจ๊ฐ ์์๋์๋ค. "
prompt += user_input
input_ids = gpt_tokenizer.encode(prompt, return_tensors="pt")
output = gpt_model.generate(input_ids, max_length=150, do_sample=True, temperature=0.8, top_k=50)
result = gpt_tokenizer.decode(output[0], skip_special_tokens=True)
return f"๐ง ๊ฐ์ ๋ถ์ ๊ฒฐ๊ณผ: {emotion}\n\nโ๏ธ GPT๊ฐ ์ด์ด ์ด ๊ธ:\n{result}"
# Gradio ์ธํฐํ์ด์ค ๊ตฌ์ฑ
gr.Interface(
fn=emotional_gpt,
inputs=gr.Textbox(lines=3, label="โ๏ธ ๊ฐ์ ์ ๋ด์ ๋ฌธ์ฅ์ ์
๋ ฅํด์ฃผ์ธ์!", placeholder="์: ์ค๋ ๋๋ฌด ์ธ๋ก์ ์ด"),
outputs="text",
title="๐ญ ๊ฐ์ ํ GPT ํ๊ธ ์๋ฌธ AI",
description="๐ง ๊ฐ์ ์ ๋จผ์ ํ์
ํ๊ณ โจ ๊ทธ ๊ฐ์ ์ ์ด์ธ๋ฆฌ๋ ๋ฌธ์ฅ์ ์ด์ด์ ์์ฑํด์ค๋๋ค!",
theme="soft",
examples=[
["๊ธฐ๋ถ์ด ๋๋ฌด ์ข์์ด"],
["์ง์ง ์ธ๋กญ๊ณ ํ๋ ํ๋ฃจ์์ด"],
["ํ์๊ฐ ๊ทธ๋ฅ ๊ทธ๋ฌ์ด"]
]
).launch()
|