shukdevdatta123 commited on
Commit
8d2d415
·
verified ·
1 Parent(s): 14010ff

Create app.py

Browse files
Files changed (1) hide show
  1. app.py +46 -0
app.py ADDED
@@ -0,0 +1,46 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import torch
2
+ import gradio as gr
3
+ from transformers import AutoModelForSequenceClassification, AutoTokenizer
4
+
5
+ # Load model and tokenizer from Hugging Face Hub
6
+ model_name = "shukdevdatta123/twitter-distilbert-base-uncased-sentiment-analysis-lora-text-classification"
7
+ model = AutoModelForSequenceClassification.from_pretrained(model_name)
8
+ tokenizer = AutoTokenizer.from_pretrained(model_name)
9
+
10
+ # Define the label mapping (binary sentiment: 0 = Negative, 1 = Positive)
11
+ id2label = {
12
+ 0: "Negative",
13
+ 1: "Positive"
14
+ }
15
+
16
+ # Set device
17
+ device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
18
+ model.to(device)
19
+
20
+ # Define the prediction function
21
+ def predict_sentiment(text):
22
+ inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True).to(device)
23
+
24
+ with torch.no_grad():
25
+ logits = model(**inputs).logits
26
+
27
+ predicted_class = torch.argmax(logits, dim=1).item()
28
+ label = id2label[predicted_class]
29
+
30
+ # Optional: add confidence score
31
+ probs = torch.nn.functional.softmax(logits, dim=1)
32
+ confidence = probs[0][predicted_class].item()
33
+
34
+ return f"{label} (Confidence: {confidence:.2f})"
35
+
36
+ # Create Gradio Interface
37
+ interface = gr.Interface(
38
+ fn=predict_sentiment,
39
+ inputs=gr.Textbox(lines=2, placeholder="Enter a sentence to analyze sentiment..."),
40
+ outputs="text",
41
+ title="Twitter Sentiment Classifier",
42
+ description="This app uses a fine-tuned DistilBERT model with LoRA adapters to predict whether a tweet or sentence is Positive or Negative."
43
+ )
44
+
45
+ # Launch the app
46
+ interface.launch()