deepseek-ai/DeepSeek-V3 ? reverse-engineered prompt

Hey, can you help me get the DeepSeek-V3 model running on my local machine? I've heard it's a really powerful new open-source AI, and I want to try chatting with it myself without using an API.

Please build me a simple Python project to run inference. It should download the main DeepSeek-V3 model from Hugging Face—not the base version—and set up all the necessary dependencies like PyTorch and transformers. The end result should be a simple script I can run from the command line. I want to be able to type in a prompt and get a text response back from the model. No need for a fancy UI. Make sure it's configured to use a GPU if available. You can look up the latest Hugging Face docs for the best way to implement this.