Spaces:
Running
Running
| import os | |
| import gradio as gr | |
| from azure.ai.inference import ChatCompletionsClient | |
| from azure.core.credentials import AzureKeyCredential | |
| # Azure Inference setup | |
| url = os.getenv("Azure_Endpoint") | |
| api_key = AzureKeyCredential(os.getenv("Azure_API_KEY")) | |
| # Initialize the ChatCompletionsClient | |
| client = ChatCompletionsClient( | |
| endpoint=url, | |
| credential=api_key, | |
| stream=True | |
| ) | |
| # Get and print model information (optional) | |
| try: | |
| model_info = client.get_model_info() | |
| print("Model name:", model_info.model_name) | |
| print("Model type:", model_info.model_type) | |
| print("Model provider name:", model_info.model_provider_name) | |
| except Exception as e: | |
| print("Could not get model info:", str(e)) | |
| # Configuration parameters | |
| default_temperature = 0.8 | |
| default_max_tokens = 2048 | |
| default_top_p = 0.1 | |
| # Example prompts that users can try | |
| example_prompts = [ | |
| "I have $20,000 in my savings account, where I receive a 4% profit per year and payments twice a year. Can you please tell me how long it will take for me to become a millionaire?", | |
| "I have total $500 create a plan with travel and food", | |
| "I have $1000 and 5 years. Is it better to invest in a stock paying $15 quarterly dividends or in a 5% annual savings account?" | |
| ] | |
| def get_azure_response(message, chat_history, temperature, max_tokens, top_p): | |
| """ | |
| Function to get a response from the Azure Phi-4 model | |
| """ | |
| # Prepare conversation history in the format expected by Azure | |
| messages = [{"role": "system", "content": "You are a helpful AI assistant."}] | |
| # Add conversation history | |
| for human, assistant in chat_history: | |
| messages.append({"role": "user", "content": human}) | |
| if assistant: # Only add non-empty assistant messages | |
| messages.append({"role": "assistant", "content": assistant}) | |
| # Add the current message | |
| messages.append({"role": "user", "content": message}) | |
| # Prepare the payload | |
| payload = { | |
| "messages": messages, | |
| "max_tokens": max_tokens, | |
| "temperature": temperature, | |
| "top_p": top_p, | |
| "presence_penalty": 0, | |
| "frequency_penalty": 0, | |
| } | |
| # Get response | |
| try: | |
| print("Sending request to Azure...") | |
| response = client.complete(payload) | |
| reply = response.choices[0].message.content | |
| # Print usage statistics | |
| print(f"Usage - Prompt tokens: {response.usage.prompt_tokens}, " | |
| f"Completion tokens: {response.usage.completion_tokens}, " | |
| f"Total tokens: {response.usage.total_tokens}") | |
| return reply | |
| except Exception as e: | |
| print(f"Error getting response: {str(e)}") | |
| return f"Error: {str(e)}" | |
| # Create the Gradio interface | |
| with gr.Blocks(title="Phi-4-mini Chatbot") as demo: | |
| gr.Markdown("Chat with the Phi-4 mini model hosted on Azure AI") | |
| # Create a chatbot component | |
| chatbot = gr.Chatbot(height=300) | |
| msg = gr.Textbox(label="Type your message here", placeholder="Ask me anything...", lines=1) | |
| clear = gr.Button("Clear Conversation") | |
| # Add examples section | |
| with gr.Accordion("Try these examples", open=True): | |
| examples = gr.Examples( | |
| examples=example_prompts, | |
| inputs=msg | |
| ) | |
| # Add model parameter controls | |
| with gr.Accordion("Model Parameters", open=False): | |
| temp_slider = gr.Slider(minimum=0.0, maximum=1.0, value=default_temperature, step=0.1, | |
| label="Temperature (higher = more creative, lower = more focused)") | |
| max_tokens_slider = gr.Slider(minimum=100, maximum=4096, value=default_max_tokens, step=100, | |
| label="Max Tokens (maximum length of response)") | |
| top_p_slider = gr.Slider(minimum=0.1, maximum=1.0, value=default_top_p, step=0.1, | |
| label="Top P (diversity of response)") | |
| # Simplified chat function that handles both sending and receiving messages | |
| def chat(message, history, temperature, max_tokens, top_p): | |
| if not message.strip(): | |
| return "", history | |
| # Get response from Azure | |
| response = get_azure_response(message, history, temperature, max_tokens, top_p) | |
| # Add the exchange to history | |
| history.append((message, response)) | |
| return "", history # Clear the input field after sending | |
| # Function to clear the conversation | |
| def clear_conversation(): | |
| return [], default_temperature, default_max_tokens, default_top_p | |
| # Set up event handlers - simplified approach | |
| msg.submit(chat, [msg, chatbot, temp_slider, max_tokens_slider, top_p_slider], [msg, chatbot]) | |
| clear.click(clear_conversation, None, [chatbot, temp_slider, max_tokens_slider, top_p_slider]) | |
| # Launch the app | |
| demo.launch(debug=True) # Set share=True to generate a public URL for testing |