| import gradio as gr | |
| from llama_cpp import Llama | |
| # Загружаем модель | |
| llm = Llama(model_path="mini_llama.gguf") | |
| def chat(message, history): | |
| output = llm(message, max_tokens=500) | |
| return output['choices'][0]['text'] | |
| # Запускаем интерфейс | |
| gr.ChatInterface( | |
| fn=chat, | |
| title="Mini Llama", | |
| description="Чат с моделью" | |
| ).launch(share=True) |