Update app.py

app.py

@@ -2,11 +2,15 @@
 
 
 
-# model = gr.load("models/microsoft/Phi-3.5-mini-instruct")
+# def generate_responce(user_input):
+#     gr.load("models/microsoft/Phi-3.5-mini-instruct")
+
+#     inputs = tokenize(user_input, return_tensor="pt")
+#     outputs = 
 
 
 # gradio_app = gr.Interface(
-#     fn=model,
+#     fn=generate_responce,
 #     inputs="text",
 #     outputs="text",
 #     max_batch_size=50,
@@ -15,61 +19,44 @@
 
 
 
-# gradio_app.launch(share=True)
-
-import gradio as gr
-
-# Load the model with gr.load()
-model_interface = gr.load("models/microsoft/Phi-3.5-mini-instruct")
-
-# Add customizations directly to the interface
-model_interface.title = "Advertisement Companion"
-model_interface.description = "Ask anything about advertising. Get professional advice and tips!"
-
-# Launch the model with batch processing
-model_interface.queue(max_size=50).launch(share=True)
-
-
+from transformers import AutoTokenizer, AutoModelForCausalLM
+import torch
 
+# Load the model and tokenizer
+tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3.5-mini-instruct", trust_remote_code=True)
+model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3.5-mini-instruct", trust_remote_code=True)
 
-# from transformers import AutoTokenizer, AutoModelForCausalLM
-# import torch
-
-# # Load the model and tokenizer
-# tokenizer = AutoTokenizer.from_pretrained("microsoft/Phi-3.5-mini-instruct", trust_remote_code=True)
-# model = AutoModelForCausalLM.from_pretrained("microsoft/Phi-3.5-mini-instruct", trust_remote_code=True)
-
-# # Define the role prompt for advertisement assistance
-# # role_prompt = "You are an advertisement assistant. Respond professionally and helpfully to advertising-related questions.\n\n"
+# Define the role prompt for advertisement assistance
+# role_prompt = "You are an advertisement assistant. Respond professionally and helpfully to advertising-related questions.\n\n"
 
-# # Function to generate responses
-# def generate_response(user_input):
-#     # Prepend role information to user input
-#     # input_text =  user_input
+# Function to generate responses
+def generate_response(user_input):
+    # Prepend role information to user input
+    # input_text =  user_input
     
-#     # Tokenize and generate response
-#     inputs = tokenizer(user_input, return_tensors="pt")
-#     outputs = model.generate(
-#         **inputs, 
-#         max_new_tokens=100,   # Increase this if you want longer responses
-#                  # Nucleus sampling to control randomness
-#     )
+    # Tokenize and generate response
+    inputs = tokenizer(user_input, return_tensors="pt")
+    outputs = model.generate(
+        **inputs, 
+        max_new_tokens=100,   # Increase this if you want longer responses
+                 # Nucleus sampling to control randomness
+    )
     
-#     # Decode and return the response
-#     response = tokenizer.decode(outputs[0], skip_special_tokens=True)
-#     return response
-
-# # Set up Gradio interface
-# interface = gr.Interface(
-#     fn=generate_response,
-#     inputs="text",
-#     outputs="text",
-#     title="Advertisement Assistant Chatbot",
-#     description="Ask me anything related to advertising. I'm here to help!"
-# )
-
-# # Launch the Gradio app with sharing enabled
-# interface.launch(share=True)
+    # Decode and return the response
+    response = tokenizer.batch_decode(outputs, skip_special_tokens=True)
+    return response
+
+# Set up Gradio interface
+interface = gr.Interface(
+    fn=generate_response,
+    inputs="text",
+    outputs="text",
+    title="Advertisement Assistant Chatbot",
+    description="Ask me anything related to advertising. I'm here to help!"
+)
+
+# Launch the Gradio app with sharing enabled
+interface.launch(share=True)
 
 
 # import gradio as gr