Update app.py

app.py

@@ -1,5 +1,10 @@
-# Load better paraphraser model
-model_name = "Vamsi/T5_Paraphrase_Paws"
+import gradio as gr
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+import torch
+
+# Choose a model: try "google/flan-t5-base" or "Vamsi/T5_Paraphrase_Paws"
+model_name = "google/flan-t5-base"
+
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
 
@@ -7,20 +12,40 @@ def humanize_text(text, tone):
     if not text.strip():
         return "⚠️ Please enter text.", "❌ Empty input"
 
-    prompt = f"paraphrase: {text} </s>"
-
-    try:
-        inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=512)
-        outputs = model.generate(
-            **inputs,
-            max_length=256,
-            num_beams=5,
-            no_repeat_ngram_size=3,
-            early_stopping=True
-        )
-        refined = tokenizer.decode(outputs[0], skip_special_tokens=True)
-        return refined.strip(), "✅ Successfully humanized!"
-    except Exception as e:
-        return f"⚠️ Error: {str(e)}", "❌ Model error"
+    # Create better prompts for FLAN-T5
+    prefix = {
+        "Academic": "Rewrite the following text in a scholarly academic tone:",
+        "Professional": "Rephrase the following text in a formal and professional style:",
+        "Conversational": "Rewrite the following text in a friendly, conversational tone:"
+    }[tone]
+
+    prompt = f"{prefix} {text}"
+
+    inputs = tokenizer(prompt, return_tensors="pt", truncation=True, max_length=512)
+    outputs = model.generate(
+        **inputs,
+        max_length=256,
+        num_beams=5,
+        no_repeat_ngram_size=3,
+        early_stopping=True
+    )
+    refined = tokenizer.decode(outputs[0], skip_special_tokens=True)
+    return refined.strip(), "✅ Successfully humanized!"
+
+iface = gr.Interface(
+    fn=humanize_text,
+    inputs=[
+        gr.Textbox(lines=8, label="✍️ Enter your text"),
+        gr.Radio(["Academic", "Professional", "Conversational"], label="Tone", value="Academic")
+    ],
+    outputs=[
+        gr.Textbox(lines=10, label="🧠 Refined Academic Rewrite"),
+        gr.Textbox(label="System Status")
+    ],
+    title="🕌 Zawiyah AI Collective – Attention-Based Humanizer AI",
+    description="Transforms ordinary text into refined, academic, or professional language using a T5 transformer model."
+)
 
+if __name__ == "__main__":
+    iface.launch()