Update app.py

app.py

@@ -1,49 +1,47 @@
 import gradio as gr
 from transformers import AutoModelForCausalLM, AutoTokenizer
 
-# Load the model and tokenizer from your Hugging Face Hub or Google Drive
-model_name = "Smilyai-labs/Smily-ultra-1"  # Replace with your model path
-
 # Load the model and tokenizer
-tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForCausalLM.from_pretrained(model_name)
-
-# Function to generate the model's response
-def chat_with_bot(user_input, chat_history):
-    # Prepare the chat history and user input
-    new_user_input_ids = tokenizer.encode(user_input + tokenizer.eos_token, return_tensors='pt')
+model = AutoModelForCausalLM.from_pretrained("Smilyai-labs/Smily-ultra-1")
+tokenizer = AutoTokenizer.from_pretrained("Smilyai-labs/Smily-ultra-1")
 
-    # Combine chat history with the new input
+# Function to generate responses from the model
+def chatbot(input_text, chat_history=[]):
+    # Encode the new user input, add the chat history as context
+    new_user_input_ids = tokenizer.encode(input_text + tokenizer.eos_token, return_tensors='pt')
+    
+    # Concatenate the chat history and the new user input
     bot_input_ids = new_user_input_ids
-    if chat_history:
-        chat_history_ids = tokenizer.encode(chat_history + tokenizer.eos_token, return_tensors='pt')
-        bot_input_ids = torch.cat([chat_history_ids, new_user_input_ids], dim=-1)
+    for history in chat_history:
+        bot_input_ids = torch.cat([history['input_ids'], bot_input_ids], dim=-1)
 
     # Generate a response from the model
-    chat_history_ids = model.generate(bot_input_ids, max_length=1024, pad_token_id=tokenizer.eos_token_id)
-
-    # Get the model's reply and update the chat history
-    bot_output = tokenizer.decode(chat_history_ids[:, bot_input_ids.shape[-1]:][0], skip_special_tokens=True)
-    chat_history += user_input + "\n" + bot_output + "\n"
+    chat_history_ids = model.generate(bot_input_ids, max_length=1000, pad_token_id=tokenizer.eos_token_id, 
+                                      temperature=0.7, top_k=50, top_p=0.95, do_sample=True, 
+                                      eos_token_id=tokenizer.eos_token_id)
     
-    return chat_history, chat_history_ids
-
-# Set up Gradio interface
-chatbot = gr.Interface(
-    fn=chat_with_bot,
-    inputs=[
-        gr.Textbox(label="User Input", placeholder="Type your message here..."),
-        gr.State(value="")
-    ],
-    outputs=[
-        gr.Textbox(label="Chat History"),
-        gr.State(value="")
-    ],
-    live=True,
-    title="Smily-ultra Chatbot",
-    description="A simple chatbot powered by the Smily-ultra model",
-    theme="compact"
-)
-
-# Launch the interface
-chatbot.launch(debug=True)
+    # Decode the generated response
+    bot_response = tokenizer.decode(chat_history_ids[0], skip_special_tokens=True)
+    
+    # Update the chat history
+    chat_history.append({'input_ids': bot_input_ids, 'response': bot_response})
+    
+    return bot_response, chat_history
+
+# Gradio interface with a more chatbot-like layout
+with gr.Blocks() as demo:
+    with gr.Column():
+        # Create a text box to display conversation history
+        chatbot_output = gr.Chatbot(label="Chatbot")
+        
+        # Create a text box for user input at the bottom
+        user_input = gr.Textbox(label="Your Message", placeholder="Type your message here...", show_label=False)
+        
+        # Create a submit button to trigger the chatbot function
+        submit_button = gr.Button("Send")
+        
+        # Link the button to the chatbot function
+        submit_button.click(chatbot, inputs=[user_input, chatbot_output], outputs=[chatbot_output, chatbot_output])
+
+# Launch the Gradio interface
+demo.launch()