import os import json import gradio as gr from llama_cpp import Llama title = "Explore the Power of WestLake-7B: Advanced Conversational AI" description = """ Welcome to the WestLake-7B Chat experience! Dive into the realm of advanced conversational AI with our cutting-edge model, WestLake-7B. This model, designed for nuanced and intelligent conversation, represents a significant leap in language processing capabilities. Explore its features and witness the future of AI-driven communication. Get your hands on the model and more details by clicking here: Westlake-7B on Hugging Face. """ llm = Llama(model_path="model.gguf", n_ctx=32768, n_threads=2, chat_format="chatml") def chat_stream_completion(message, history, system_prompt): messages_prompts = [{"role": "system", "content": system_prompt}] for human, assistant in history: messages_prompts.append({"role": "user", "content": human}) messages_prompts.append({"role": "assistant", "content": assistant}) messages_prompts.append({"role": "user", "content": message}) response = llm.create_chat_completion( messages=messages_prompts, stream=True ) message_repl = "" for chunk in response: if len(chunk['choices'][0]["delta"]) != 0 and "content" in chunk['choices'][0]["delta"]: message_repl = message_repl + chunk['choices'][0]["delta"]["content"] yield message_repl gr.ChatInterface( fn=chat_stream_completion, title=title, description=description, additional_inputs=[gr.Textbox("Perform the task to the best of your ability.")], additional_inputs_accordion="📝 System prompt", examples=[ ["Summarize the key themes of '1984' by George Orwell."], ["Convert 150 Fahrenheit to Celsius."], ["Generate a short story about a time-traveling detective."], ["Explain the concept of black holes in simple terms."], ["What are the health benefits of Mediterranean diet?"], ["Write Python code to sort a list of numbers in ascending order."], ["Translate 'Hello, how are you?' into French."], ["Create a motivational quote about perseverance."] ] ).queue().launch(server_name="0.0.0.0")