Update app.py

app.py

@@ -1,10 +1,9 @@
 import gradio as gr
-from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
+from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
 
-# ✅ Your fine-tuned model on Hugging Face
 MODEL_NAME = "Abeersherif/Medical_Homework2"
 
-# Load model & tokenizer from HF
+# Load model and tokenizer
 tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
 model = AutoModelForCausalLM.from_pretrained(MODEL_NAME)
 
@@ -15,37 +14,40 @@ pipe = pipeline(
 )
 
 def respond(message, history, system_message, max_tokens, temperature, top_p):
-    # Combine chat history manually (since no chat template exists)
-    full_prompt = f"System: {system_message}\n"
-    for turn in history:
-        role = turn["role"]
-        text = turn["content"]
-        full_prompt += f"{role.capitalize()}: {text}\n"
-
-    full_prompt += f"User: {message}\nAssistant:"
-
-    out = pipe(
-        full_prompt,
-        max_new_tokens=max_tokens,
-        temperature=temperature,
-        top_p=top_p,
+    """
+    message: str                    -> latest user message
+    history: list[[user, bot], ...] -> previous chat turns (default tuples mode)
+    """
+
+    # Build a simple text prompt (no chat template)
+    conversation = f"System: {system_message}\n\n"
+    for user_msg, bot_msg in history:
+        conversation += f"User: {user_msg}\nAssistant: {bot_msg}\n"
+    conversation += f"User: {message}\nAssistant:"
+
+    result = pipe(
+        conversation,
+        max_new_tokens=int(max_tokens),
+        temperature=float(temperature),
+        top_p=float(top_p),
         do_sample=True,
     )[0]["generated_text"]
 
-    # return only assistant after last "Assistant:"
-    if "Assistant:" in out:
-        out = out.split("Assistant:", 1)[-1]
-
-    return out.strip()
+    # Keep only what the assistant said last
+    if "Assistant:" in result:
+        result = result.split("Assistant:")[-1]
 
+    return result.strip()
 
 
 chatbot = gr.ChatInterface(
     fn=respond,
-    type="messages",
+    # ⚠️ use default history format (tuples), do NOT set type="messages" here
     additional_inputs=[
-        gr.Textbox("You are a helpful medical assistant. Answer concisely with brief reasoning.",
-                   label="System message"),
+        gr.Textbox(
+            "You are a helpful medical assistant. Answer concisely with brief reasoning.",
+            label="System message",
+        ),
         gr.Slider(1, 512, value=256, step=1, label="Max new tokens"),
         gr.Slider(0.1, 4.0, value=0.7, step=0.1, label="Temperature"),
         gr.Slider(0.1, 1.0, value=0.9, step=0.05, label="Top-p"),