Spaces:
Running
Running
Update emotion_detection.py
Browse files- emotion_detection.py +35 -25
emotion_detection.py
CHANGED
@@ -1,13 +1,11 @@
|
|
1 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
2 |
from transformers_interpret import SequenceClassificationExplainer
|
3 |
import torch
|
4 |
-
import pandas as pd
|
5 |
|
6 |
|
7 |
class EmotionDetection:
|
8 |
"""
|
9 |
Emotion Detection on text data.
|
10 |
-
|
11 |
Attributes:
|
12 |
tokenizer: An instance of Hugging Face Tokenizer
|
13 |
model: An instance of Hugging Face Model
|
@@ -20,55 +18,67 @@ class EmotionDetection:
|
|
20 |
self.model = AutoModelForSequenceClassification.from_pretrained(hub_location)
|
21 |
self.explainer = SequenceClassificationExplainer(self.model, self.tokenizer)
|
22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
23 |
def justify(self, text):
|
24 |
"""
|
25 |
Get html annotation for displaying emotion justification over text.
|
26 |
-
|
27 |
Parameters:
|
28 |
-
text (str): The user input string
|
29 |
-
|
30 |
Returns:
|
31 |
-
html (
|
32 |
"""
|
33 |
-
|
34 |
word_attributions = self.explainer(text)
|
35 |
-
html = self.explainer.visualize(
|
36 |
-
|
37 |
return html
|
38 |
|
39 |
def classify(self, text):
|
40 |
"""
|
41 |
-
Recognize Emotion in text.
|
42 |
-
|
43 |
Parameters:
|
44 |
text (str): The user input string to perform emotion classification on
|
45 |
-
|
46 |
Returns:
|
47 |
-
|
48 |
"""
|
49 |
-
|
50 |
-
tokens = self.tokenizer.encode_plus(text, add_special_tokens=False, return_tensors='pt')
|
51 |
outputs = self.model(**tokens)
|
52 |
probs = torch.nn.functional.softmax(outputs[0], dim=-1)
|
53 |
probs = probs.mean(dim=0).detach().numpy()
|
|
|
54 |
labels = list(self.model.config.id2label.values())
|
55 |
-
|
|
|
|
|
56 |
|
57 |
-
|
|
|
|
|
|
|
|
|
58 |
|
59 |
def run(self, text):
|
60 |
"""
|
61 |
Classify and Justify Emotion in text.
|
62 |
-
|
63 |
Parameters:
|
64 |
text (str): The user input string to perform emotion classification on
|
65 |
-
|
66 |
Returns:
|
67 |
-
|
68 |
-
html (
|
69 |
"""
|
70 |
-
|
71 |
-
preds = self.classify(text)
|
72 |
html = self.justify(text)
|
73 |
-
|
74 |
-
return preds, html
|
|
|
1 |
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
2 |
from transformers_interpret import SequenceClassificationExplainer
|
3 |
import torch
|
|
|
4 |
|
5 |
|
6 |
class EmotionDetection:
|
7 |
"""
|
8 |
Emotion Detection on text data.
|
|
|
9 |
Attributes:
|
10 |
tokenizer: An instance of Hugging Face Tokenizer
|
11 |
model: An instance of Hugging Face Model
|
|
|
18 |
self.model = AutoModelForSequenceClassification.from_pretrained(hub_location)
|
19 |
self.explainer = SequenceClassificationExplainer(self.model, self.tokenizer)
|
20 |
|
21 |
+
# Friendly emoji mapping
|
22 |
+
self.emoji_map = {
|
23 |
+
"joy": "π",
|
24 |
+
"anger": "π ",
|
25 |
+
"optimism": "π",
|
26 |
+
"sadness": "π’"
|
27 |
+
}
|
28 |
+
|
29 |
+
# Friendly explanation mapping
|
30 |
+
self.explanation_map = {
|
31 |
+
"joy": "The person is happy or excited.",
|
32 |
+
"anger": "The person is upset or angry.",
|
33 |
+
"optimism": "The person is feeling hopeful or positive.",
|
34 |
+
"sadness": "The person is feeling low or unhappy."
|
35 |
+
}
|
36 |
+
|
37 |
def justify(self, text):
|
38 |
"""
|
39 |
Get html annotation for displaying emotion justification over text.
|
|
|
40 |
Parameters:
|
41 |
+
text (str): The user input string for emotion justification
|
|
|
42 |
Returns:
|
43 |
+
html (str): html string for plotting emotion prediction justification
|
44 |
"""
|
|
|
45 |
word_attributions = self.explainer(text)
|
46 |
+
html = self.explainer.visualize(return_html=True) # Changed to return HTML string
|
|
|
47 |
return html
|
48 |
|
49 |
def classify(self, text):
|
50 |
"""
|
51 |
+
Recognize Emotion in text (simplified).
|
|
|
52 |
Parameters:
|
53 |
text (str): The user input string to perform emotion classification on
|
|
|
54 |
Returns:
|
55 |
+
result (str): User-friendly emotion label with emoji and explanation
|
56 |
"""
|
57 |
+
tokens = self.tokenizer.encode_plus(text, return_tensors='pt')
|
|
|
58 |
outputs = self.model(**tokens)
|
59 |
probs = torch.nn.functional.softmax(outputs[0], dim=-1)
|
60 |
probs = probs.mean(dim=0).detach().numpy()
|
61 |
+
|
62 |
labels = list(self.model.config.id2label.values())
|
63 |
+
max_index = probs.argmax()
|
64 |
+
emotion = labels[max_index]
|
65 |
+
confidence = probs[max_index]
|
66 |
|
67 |
+
emoji = self.emoji_map.get(emotion, "")
|
68 |
+
explanation = self.explanation_map.get(emotion, "")
|
69 |
+
|
70 |
+
result = f"{emoji} **{emotion.capitalize()}** ({confidence:.1%})\n{explanation}"
|
71 |
+
return result
|
72 |
|
73 |
def run(self, text):
|
74 |
"""
|
75 |
Classify and Justify Emotion in text.
|
|
|
76 |
Parameters:
|
77 |
text (str): The user input string to perform emotion classification on
|
|
|
78 |
Returns:
|
79 |
+
result (str): Friendly emotion classification output
|
80 |
+
html (str): HTML visualization string for model justification
|
81 |
"""
|
82 |
+
result = self.classify(text)
|
|
|
83 |
html = self.justify(text)
|
84 |
+
return result, html
|
|