Update app.py

app.py

@@ -1,23 +1,24 @@
 import gradio as gr
 import pandas as pd
 import torch
-from transformers import pipeline
+import transformers
 
 # Load the model pipeline
-model_id = "meta-llama/Llama-3.2-1B"
-pipe = pipeline(
-    "text-generation", 
-    model=model_id, 
-    torch_dtype=torch.bfloat16, 
-    device_map="auto"
+model_id = "meta-llama/Meta-Llama-3.1-8B-Instruct"
+pipeline = transformers.pipeline(
+    "text-generation",
+    model=model_id,
+    model_kwargs={"torch_dtype": torch.bfloat16},
+    device_map="auto",
 )
 
-# Define the system message for the model
-system_message = (
-    "You are an experienced Fashion designer who starts conversation with proper greeting, "
-    "giving valuable and catchy fashion advice and suggestions, stays to the point and precise, "
-    "asks questions only if the user has any concerns over your provided suggestions."
-)
+# Define the initial system message
+system_message = {
+    "role": "system",
+    "content": "You are an experienced Fashion designer who starts conversation with proper greeting, "
+               "giving valuable and catchy fashion advice and suggestions, stays to the point and precise, "
+               "asks questions only if the user has any concerns over your provided suggestions."
+}
 
 # Function to reset the chat
 def reset_chat():
@@ -54,19 +55,23 @@ def chat(user_input, messages):
         messages.append({"role": "user", "content": user_input})
 
         # Prepare the input for the model
-        input_text = system_message + "\n" + "\n".join([f"{msg['role']}: {msg['content']}" for msg in messages])
-        
+        input_text = messages.copy()  # Make a copy of messages
+
+        # Generate a response using the pipeline
         try:
-            # Generate a response using the pipeline
-            response = pipe(input_text, max_length=150, num_return_sequences=1, temperature=0.7)
-            response_content = response[0]['generated_text'].split('\n')[-1].strip()  # Extract the last line of the generated text
+            # Convert the messages to a format the model can understand
+            formatted_input = "\n".join([f"{msg['role']}: {msg['content']}" for msg in input_text])
+            response = pipeline(formatted_input, max_new_tokens=256)
+
+            # Extract the assistant's response
+            response_content = response[0]["generated_text"].strip()
+            
+            # Store assistant response in the chat history
+            messages.append({"role": "assistant", "content": response_content})
 
         except Exception as e:
             response_content = f"Error: {str(e)}"
 
-        # Store assistant response in the chat history
-        messages.append({"role": "assistant", "content": response_content})
-
         return messages, response_content
     return messages, ""