Update app.py

app.py

@@ -1,24 +1,23 @@
 import gradio as gr
 import torch
-from modeling import BERTMultiLabel
 from transformers import AutoTokenizer
-
+from modeling import BERTMultiLabel
 
 LABELS = ["anger", "fear", "joy", "sadness", "surprise"]
 
-# ---- Load tokenizer ----
+# Load tokenizer and model
 tokenizer = AutoTokenizer.from_pretrained("./")
+model = BERTMultiLabel("microsoft/deberta-v3-base", num_labels=len(LABELS))
 
-
-model = BERTMultiLabel(model_name="microsoft/deberta-v3-base", num_labels=len(LABELS))
 state = torch.load("pytorch_model.bin", map_location="cpu")
 model.load_state_dict(state)
 model.eval()
 
 
+# ---------- PREDICTION FUNCTION ----------
 def predict(text):
-    if text.strip() == "":
-        return {"error": "Please enter text"}
+    if not text.strip():
+        return {"error": "Please enter text."}
 
     enc = tokenizer(
         text,
@@ -30,39 +29,124 @@ def predict(text):
 
     with torch.no_grad():
         logits = model(enc["input_ids"], enc["attention_mask"])
-        probs = torch.sigmoid(logits).flatten().tolist()
+        probs = torch.sigmoid(logits)[0].tolist()
+
+    scores = {label: round(p, 4) for label, p in zip(LABELS, probs)}
+    main_emotion = LABELS[int(torch.tensor(probs).argmax())]
 
-    result = {label: float(f"{p:.4f}") for label, p in zip(LABELS, probs)}
-    top_label = LABELS[int(torch.argmax(torch.tensor(probs)))]
+    emoji_map = {
+        "anger": "😡",
+        "fear": "😨",
+        "joy": "😊",
+        "sadness": "😢",
+        "surprise": "😮",
+    }
 
     return {
-        "Predicted Mood": top_label,
-        "Scores": result
+        "Predicted Mood": f"{emoji_map[main_emotion]} {main_emotion.capitalize()}",
+        "Scores": scores
     }
 
 
-with gr.Blocks(title="Mood Detector") as app:
-    gr.Markdown("# 🧠 DeBERTa Mood Detector")
-    gr.Markdown("Enter any text to analyze the emotional tone.")
+# ---------- UI LAYOUT ----------
+with gr.Blocks(
+    title="Sentiment & Emotion Analyzer",
+    theme=gr.themes.Soft(
+        primary_hue="indigo",
+        secondary_hue="blue",
+        radius="lg",
+        text_size="lg"
+    )
+) as demo:
+
+    # ---- HEADER ----
+    gr.Markdown(
+        """
+        <div style="text-align:center; padding: 20px;">
+            <h1 style="font-size: 3rem; margin-bottom: 0;">🎭 Multi-Label Emotion Detection</h1>
+            <p style="font-size:1.2rem; color:#666;">
+                Powered by DeBERTa-v3 • Fine-tuned on IIT Madras Deep Learning Competition Dataset
+            </p>
+        </div>
+        """
+    )
 
     with gr.Row():
-        # LEFT PANEL
+
+        # ---------- LEFT PANEL (FULL INFO) ----------
         with gr.Column(scale=1):
-            gr.Markdown("""
-            ## 📌 Model Info
-            - **Architecture:** DeBERTa-v3 Base  
-            - **Task:** Multi-label emotion classification  
-            - **Trained by:** You  
-            - **Dataset:** Custom text emotion dataset  
-            - **Output:** anger, fear, joy, sadness, surprise  
-            """)
-        
-        # RIGHT PANEL
+
+            gr.Markdown(
+                """
+                <div style="
+                    background:white; border-radius:16px; padding:20px;
+                    box-shadow:0 4px 20px rgba(0,0,0,0.08); margin-bottom:20px;
+                ">
+                    <h2>📌 Model Overview</h2>
+                    <ul style="line-height:1.6;">
+                        <li><b>Architecture:</b> DeBERTa-v3 Base</li>
+                        <li><b>Task:</b> Multi-Label Emotion Classification</li>
+                        <li><b>Labels:</b> anger, fear, joy, sadness, surprise</li>
+                        <li><b>Training Framework:</b> PyTorch + HuggingFace</li>
+                        <li><b>Loss:</b> BCEWithLogitsLoss</li>
+                        <li><b>Optimizer:</b> AdamW (LR = 1e-5)</li>
+                    </ul>
+                </div>
+
+                <div style="
+                    background:white; border-radius:16px; padding:20px;
+                    box-shadow:0 4px 20px rgba(0,0,0,0.08); margin-bottom:20px;
+                ">
+                    <h2>📊 Dataset Details</h2>
+                    <p style="line-height:1.6;">
+                        Dataset used in IIT Madras Deep Learning & GenAI (Sep 2025).
+                        Multi-label emotion classification with 5 emotions:
+                    </p>
+                    <ul>
+                        <li>😠 Anger</li>
+                        <li>😨 Fear</li>
+                        <li>😊 Joy</li>
+                        <li>😢 Sadness</li>
+                        <li>😲 Surprise</li>
+                    </ul>
+                    <p><b>Evaluation Metric:</b> Macro F1-Score</p>
+                </div>
+
+                <div style="
+                    background:white; border-radius:16px; padding:20px;
+                    box-shadow:0 4px 20px rgba(0,0,0,0.08);
+                ">
+                    <h2>🏆 Project Highlights</h2>
+                    <ul style="line-height:1.6;">
+                        <li>Competition Rank: <b>27 / 200</b></li>
+                        <li>Public F1 Score: <b>87.80%</b></li>
+                        <li>Private F1 Score: <b>87.00%</b></li>
+                        <li>Experimented with CNN, GRU, BiLSTM, DistilBERT & DeBERTa</li>
+                    </ul>
+                </div>
+                """
+            )
+
+        # ---------- RIGHT PANEL (ANALYZER) ----------
         with gr.Column(scale=2):
-            user_input = gr.Textbox(label="Enter text here", placeholder="Type something like: 'I feel so happy today!'")
-            analyze_button = gr.Button("Analyze Mood")
-            output = gr.JSON(label="Result")
 
-            analyze_button.click(predict, inputs=user_input, outputs=output)
+            with gr.Group():
+                user_input = gr.Textbox(
+                    label="Enter your text",
+                    placeholder="Example: I feel amazing today! 🎉",
+                    lines=4
+                )
+                analyze_btn = gr.Button("🎯 Analyze Emotion", size="lg")
+                output = gr.JSON(label="Model Output")
+
+            analyze_btn.click(predict, inputs=user_input, outputs=output)
+
+    gr.Markdown(
+        """
+        <br><p style="text-align:center; color:#888;">
+        Built by <b>Ayusman Samasi</b> • IIT Madras Deep Learning & GenAI
+        </p>
+        """
+    )
 
-app.launch()
+demo.launch()