Update app.py

app.py

@@ -4,28 +4,50 @@ import torch
 import numpy as np
 
 # Load model and tokenizer
-model_name = "KevSun/IELTS_essay_scoring"
+model_name = "JacobLinCool/IELTS_essay_scoring_safetensors"
 model = AutoModelForSequenceClassification.from_pretrained(model_name)
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 
 # Prediction function
 def score_essay(essay):
-    inputs = tokenizer(essay, return_tensors="pt", truncation=True, max_length=512)
+    if not essay.strip():
+        return {"Task Achievement": 0,
+                "Coherence & Cohesion": 0,
+                "Vocabulary": 0,
+                "Grammar": 0,
+                "Overall": 0}
+
+    # Tokenize and truncate to max 512 tokens
+    inputs = tokenizer(
+        essay,
+        return_tensors="pt",
+        truncation=True,
+        max_length=512
+    )
+
+    # Run inference
     with torch.no_grad():
         outputs = model(**inputs)
+
+    # Extract logits and normalize to 9
     preds = outputs.logits.squeeze().numpy()
     normalized = (preds / preds.max()) * 9
     rounded = np.round(normalized * 2) / 2
+
+    # Map labels
     labels = ["Task Achievement", "Coherence & Cohesion", "Vocabulary", "Grammar", "Overall"]
     return {label: float(score) for label, score in zip(labels, rounded)}
 
 # Gradio UI
-iface = gr.Interface(
-    fn=score_essay,
-    inputs=gr.Textbox(lines=10, placeholder="Paste your IELTS essay here..."),
-    outputs=[gr.Label(num_top_classes=5)],
-    title="Automated IELTS Essay Scorer",
-    description="Predicts scores for multiple dimensions of your essay"
-)
-
-iface.launch()
+with gr.Blocks() as demo:
+    gr.Markdown("## Automated IELTS Essay Scoring")
+    gr.Markdown("Paste your essay below to get scores for all dimensions (Task, Coherence, Vocabulary, Grammar, Overall).")
+
+    essay_input = gr.Textbox(lines=10, placeholder="Paste your IELTS essay here...")
+    score_output = gr.Label()
+
+    submit_btn = gr.Button("Score Essay")
+    submit_btn.click(fn=score_essay, inputs=essay_input, outputs=score_output)
+
+# Launch app (for HF Spaces, leave default)
+demo.launch()