Spaces:
Sleeping
Sleeping
Upload 2 files
Browse files- app.py +42 -0
- sentiment7_model_acc0.9653.pth +3 -0
app.py
ADDED
|
@@ -0,0 +1,42 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
import torch
|
| 3 |
+
from transformers import BertConfig, BertForSequenceClassification, BertTokenizer
|
| 4 |
+
import numpy as np
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
# Load the model and tokenizer
|
| 8 |
+
def load_model():
|
| 9 |
+
tokenizer = BertTokenizer.from_pretrained('beomi/kcbert-base')
|
| 10 |
+
config = BertConfig.from_pretrained('beomi/kcbert-base', num_labels=7)
|
| 11 |
+
model = BertForSequenceClassification.from_pretrained('beomi/kcbert-base', config=config)
|
| 12 |
+
model_state_dict = torch.load('sentiment7_model_acc8878.pth', map_location=torch.device('cpu')) # cpu 사용
|
| 13 |
+
model.load_state_dict(model_state_dict)
|
| 14 |
+
model.eval()
|
| 15 |
+
return model, tokenizer
|
| 16 |
+
|
| 17 |
+
model, tokenizer = load_model()
|
| 18 |
+
|
| 19 |
+
# Define the inference function
|
| 20 |
+
def inference(input_doc):
|
| 21 |
+
inputs = tokenizer(input_doc, return_tensors='pt')
|
| 22 |
+
outputs = model(**inputs)
|
| 23 |
+
probs = torch.softmax(outputs.logits, dim=1).squeeze().tolist()
|
| 24 |
+
class_idx = {'공포': 0, '놀람': 1, '분노': 2, '슬픔': 3, '중립': 4, '행복': 5, '혐오': 6}
|
| 25 |
+
results = {class_name: prob for class_name, prob in zip(class_idx, probs)}
|
| 26 |
+
# Find the class with the highest probability
|
| 27 |
+
max_prob_class = max(results, key=results.get)
|
| 28 |
+
max_prob = results[max_prob_class]
|
| 29 |
+
# Display results
|
| 30 |
+
return [results, f"가장 강하게 나타난 감정: {max_prob_class}"]
|
| 31 |
+
''' for class_name, prob in results.items():
|
| 32 |
+
print(f"{class_name}: {prob:.2%}")'''
|
| 33 |
+
|
| 34 |
+
# Set up the Streamlit interface
|
| 35 |
+
st.title('감정분석(Sentiment Analysis)')
|
| 36 |
+
st.markdown('<small style="color:grey;">글에 나타난 공포, 놀람, 분노, 슬픔, 중립, 행복, 혐오의 정도를 비율로 알려드립니다.</small>', unsafe_allow_html=True)
|
| 37 |
+
user_input = st.text_area("이 곳에 글 입력(100자 이하 권장):")
|
| 38 |
+
if st.button('시작'):
|
| 39 |
+
result = inference(user_input)
|
| 40 |
+
st.write(result[0])
|
| 41 |
+
st.write(result[1])
|
| 42 |
+
|
sentiment7_model_acc0.9653.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3a6e54803c4c1d0eae2c24cc17cf411f68f8e5e37af89dce068b3140abc7761d
|
| 3 |
+
size 435781207
|