Update app.py

app.py

@@ -1,34 +1,21 @@
-import torch
-from transformers import AutoTokenizer, AutoModelForQuestionAnswering
 import gradio as gr
+import torch
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
 
-def answer_question(question, context):
-    # Load the model and tokenizer
-    model_name = "Qwen/Qwen2-VL"
-    tokenizer = AutoTokenizer.from_pretrained(model_name)
-    model = AutoModelForQuestionAnswering.from_pretrained(model_name)
-
-    # Tokenize the input
-    inputs = tokenizer(question, context, return_tensors="pt")
-
-    # Generate the answer
+def classify_text(text):
+    tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
+    model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
+    inputs = tokenizer(text, return_tensors="pt")
     with torch.no_grad():
         outputs = model(**inputs)
-
-    # Decode the answer
-    answer = tokenizer.decode(outputs.start_logits.argmax(), skip_special_tokens=True)
-
-    return answer
+    predicted_class = torch.argmax(outputs.logits, dim=1).item()
+    return "Positive" if predicted_class == 1 else "Negative"
 
 interface = gr.Interface(
-    fn=answer_question,
-    inputs=[
-        gr.inputs.Textbox(label="Question"),
-        gr.inputs.Textbox(label="Context")
-    ],
+    fn=classify_text,
+    inputs=gr.inputs.Textbox(label="Text"),
     outputs="text",
-    title="Qwen2-VL Question Answering",
-    description="Enter a question and context to get an answer.",
+    title="Sentiment Analysis",
+    description="Enter a text to classify its sentiment as positive or negative.",
 )
-
 interface.launch()