Update app.py

app.py

@@ -0,0 +1,104 @@
+import gradio as gr
+import torch
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+import torch.nn.functional as F
+
+# Lightweight model already fine-tuned for sentiment (91%+ accuracy)
+# Much smaller than bert-base-uncased → fits easily in free CPU Spaces
+MODEL_NAME = "distilbert/distilbert-base-uncased-finetuned-sst-2-english"
+
+# Load model efficiently
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+
+model = AutoModelForSequenceClassification.from_pretrained(
+    MODEL_NAME,
+    torch_dtype=torch.float32,   # Safe for CPU
+    low_cpu_mem_usage=True
+)
+
+model.eval()
+
+# Safe device handling
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model = model.to(device)
+
+def predict_sentiment(text: str):
+    if not text or not text.strip():
+        return "Please enter some text", 0.0, "⚠️"
+
+    # Tokenize
+    inputs = tokenizer(
+        text, 
+        return_tensors="pt", 
+        truncation=True, 
+        padding=True, 
+        max_length=512
+    ).to(device)
+
+    # Inference with no gradient
+    with torch.no_grad():
+        outputs = model(**inputs)
+        probs = F.softmax(outputs.logits, dim=-1)
+
+    # Get prediction and confidence
+    pred = torch.argmax(probs, dim=-1).item()
+    confidence = probs[0][pred].item() * 100
+
+    if pred == 1:  # 1 = positive in this model
+        sentiment = "Positive 😊"
+        emoji = "🟢"
+    else:
+        sentiment = "Negative 😞"
+        emoji = "🔴"
+
+    return sentiment, round(confidence, 2), emoji
+
+# ====================== Gradio UI ======================
+with gr.Blocks(theme=gr.themes.Soft(), title="BERT Sentiment Analysis") as demo:
+    gr.Markdown("# 🎬 Transformer Sentiment Analysis")
+    gr.Markdown("**DistilBERT** (lightweight BERT) for fast movie review / text sentiment prediction")
+
+    with gr.Row():
+        text_input = gr.Textbox(
+            label="Enter your text (movie review, comment, tweet, etc.)",
+            placeholder="Type or paste here...",
+            lines=5,
+        )
+
+    analyze_btn = gr.Button("🔍 Analyze Sentiment", variant="primary", size="large")
+
+    with gr.Row():
+        sentiment_output = gr.Textbox(label="Sentiment Result", interactive=False)
+        confidence_output = gr.Number(label="Confidence (%)", interactive=False)
+        indicator_output = gr.Textbox(label="Indicator", interactive=False)
+
+    # Nice examples
+    gr.Examples(
+        examples=[
+            ["This movie was absolutely fantastic! The acting and story were brilliant."],
+            ["I really disliked the plot. It was boring and predictable."],
+            ["The visuals were great but the dialogue felt terrible."],
+            ["One of the best films I have watched this year. Highly recommended!"],
+            ["Complete waste of time. Do not watch this movie."],
+        ],
+        inputs=text_input,
+        outputs=[sentiment_output, confidence_output, indicator_output],
+        fn=predict_sentiment,
+        cache_examples=False,
+    )
+
+    # Button click
+    analyze_btn.click(
+        fn=predict_sentiment,
+        inputs=text_input,
+        outputs=[sentiment_output, confidence_output, indicator_output],
+    )
+
+# Launch (Important: share=False on HF Spaces)
+if __name__ == "__main__":
+    demo.launch(
+        server_name="0.0.0.0",
+        server_port=7860,
+        share=False,
+        debug=False
+    )