DeepSeek Registration and Local Model Deployment Guide

围绕「DeepSeek Registration and Local Model Deployment Guide」，本文保留关键参数与套餐要点，并补充选购时容易忽略的细节。

I. Register and Use DeepSeek Online

1. Visit the Official Website:

   • Open the DeepSeek official website: https://www.deepseek.com/

2. Register an Account:

   • Click the “Register” or “Sign Up” button on the page
   • Enter your phone number, set a password, and follow the prompts to complete registration
   • If you need a Chinese phone number, you can use an SMS activation platform for assisted registration. Below are the registration channels:
   • Quick Registration Link for the China Version of SMS Activation Platform - Direct Access to DeepSeek
   • Quick Registration Link for the International Version of SMS Activation Platform - Direct Access to DeepSeek

II. Local DeepSeek Model Deployment

1. Install Ollama:

   • Ollama is a tool for managing and running large language models locally
   • Visit the Ollama official website
   • Download and install the appropriate version for your operating system

2. Download DeepSeek Model:

   • In your terminal or command line, use Ollama to download the DeepSeek-R1 model:

     ollama pull deepseek-r1
     

   • Wait for the model download to complete

3. Run the Model:

   • Enter the following command in your terminal to start the model:

     ollama run deepseek-r1
     

   • Once the model starts, you can interact with it in the terminal by entering your questions, and the model will generate corresponding responses

4. Visual Interface (Optional):

   • If you prefer using a graphical interface to interact with the model, you can install Chatbox
   • Visit the Chatbox project page
   • Download and install the version suitable for your system
   • Configure Chatbox to use your local DeepSeek model, and you can interact through the graphical interface

• Hardware Requirements: Running large models requires high hardware specifications. For example, a 1.5B parameter model requires at least a GPU with 4GB VRAM and 16GB RAM. If hardware specifications are insufficient, the model may use CPU for computation, which will result in longer processing times.

• Model Size: DeepSeek-R1 model comes in different parameter sizes. For example, the 7B parameter model is approximately 4.7GB. Choose the appropriate model version based on your hardware configuration.