5 Minutes on deepseek Localization Deployment

Step 1: Install a "magic tool" Ollama 🚀 (Windows computers look here!)

• What's Ollama? 🤔 As a side note, Ollama is a "magic toolkit" that makes it easy to run all kinds of awesome AI models like DeepSeek, which we're going to be using today, and which is like a computer game that you can play with a single click!
• How do I install Ollama on Windows? Super easy! There are two ways to do it, the graphical interface and the command line, and I recommend you use theGraphical User Interface (GUI) installationIt's just a matter of pointing and clicking, just like installing normal software!
  

  🌟 Method 1: GUI Installation (Recommended!) 🌟

  • Download the installation package: Open your browser and visit Ollama Official website: https://ollama.com/ You will see a large Download for Windows button and be brave enough to click it! 🖱️
    

    💡 Tip: If the official website doesn't open, try a domestic mirror site or use the Magic Internet Tool.

  • Run the installation program: Once the download is complete, find your downloaded .exe Installation file, double-click to run it. Just like the usual installation of QQ, WeChat, all the way to "Next" on OK! The installation process is very simple and you don't have to worry about it. 😌
    

    📌 Note: During the installation process, some security tips may pop up, don't worry, just choose "Allow" or "Yes".

  • Installation complete! Once the installation is complete, Ollama will be installed on your Windows computer! 🎉 Wasn't it easy?

  🛠️ Method 2: Command Line Installation (for those of you who like to tinker) 🛠️

  If you are a command line expert or prefer a more "technical" installation, you can also install Ollama from the command line.

  • Open PowerShell: In the Windows search bar, type PowerShell , find "Windows PowerShell", right click on it and select "Run as administrator".
  • Copy and paste commands: In a PowerShell window, copy and paste the following line of code and press Enter:

    curl -fsSL https://ollama.com/install.ps1 | iex
    

    💡 Tip: This line of code is similar to the incantation on Linux/macOS that automatically installs Ollama for you. powerShell will automatically download and run the installation script.

  • Wait for the installation to complete: PowerShell will display the installation process, so be patient for a while until the installation is complete.

  🚀 Verify Installation (common to Windows & other systems)

  Regardless of which method you used to install Ollama, once the installation is complete, it is recommended to verify that the installation was successful. Open yourCommand Prompt (CMD) or PowerShell (Windows), orTerminal (macOS/Linux), type the following line of command and press Enter:

  ollama --version
  

  If you see the Ollama version number displayed, you have successfully installed Ollama! 🎉 Congratulations on taking the first step in your AI journey!

Step 2: Equip Ollama with a "smart brain" DeepSeek 🧠

• What's DeepSeek? 🤓 DeepSeek The "Smart Brain" we're going to use today is an awesome AI model that can talk to you, write articles, write code, and simply be your all-around assistant!
• How do I load DeepSeek into Ollama? It's still very simple, just add a few words to yourCommand Prompt (CMD) or PowerShell (Windows), orTerminal (macOS/Linux), enter the following line of code:

  ollama run deepseek-r1:8b
  

  📌 Pay attention:deepseek-r1:8b This is the same "small" DeepSeek model for the first time. If you have a good computer, or want to play with more powerful models:DeepSeek R1 671B Local Deployment Tutorial: Based on Ollama and Dynamic Quantization

  Once you run this line of command, Ollama will automatically download the DeepSeek model. Wait patiently for the download to complete... ⏳

Recommended Reading

What are the minimum hardware requirements if deploying the deepseek model locally?

Private Deployment without Local GPUs DeepSeek-R1 32B(Here's a tutorial on how to install Ollama the other way around)

DeepSeek Official Pick: A Guide to Practical AI Tools with DeepSeek R1 Integration

Avoid the pit guide: Taobao DeepSeek R1 installation package paid upsell? Teach you local deployment for free (with one-click installer)

Step 3: Start chatting with DeepSeek! 🎉

• Run command: If the model download is successful, Ollama will automatically launch DeepSeek, and you will see a prompt in your terminal telling you that DeepSeek is ready to chat with you!
• Start a dialog: Now, directly in theCommand Prompt (CMD) or PowerShell (Windows), orTerminal (macOS/Linux), type in the question you want to ask DeepSeek, or the thing you want it to do, for example:

  你好 DeepSeek！
  

  Press enter and DeepSeek will respond to you like a real person! You can ask it about the weather, ask it about history, have it write stories, code... explore the magic of AI! ✨

  DeepSeek: 你好！很高兴认识你！有什么我可以帮你的吗？
  

  💬 Another chestnut: you can ask DeepSeek "What to do if I'm having a bad day?" , or "Write me a poem about the stars" and see what it surprises you with!

Model selection and "diet" version 💡

• Model parameters: remember that 8b , 14b These numbers? They represent the "smartness" of the model. The bigger the number, the more powerful the model, but also the "heavier" it is, and the more demanding it is on the computer.
• Quantitative version: "Quantization" is like "slimming down" the models so that they are lighter and run smoothly on a normal computer. If you have an average computer or want to try a bigger and more powerful model, consider the quantized version.
  

  💡 Version Selection Suggestion:

  • First time experience / Average computer configuration: deepseek-r1:8b or deepseek-r1:14b are good choices, fast speed and less resource consumption.
  • Want to experience a stronger DeepSeek: You can try. deepseek-r1:32b Or a larger version, but it may require a better computer configuration.
  • Quantitative version: If you want to run large models on an average configured computer, try the quantized version mentioned in the article DeepSeek-R1-Distill-Qwen-32B-GGUF The install command is a little more complicated, but the results are great!

Notes for Windows Users ⚠️

• Antivirus software: Some antivirus programs may misdiagnose the Ollama installer, if you encounter this, you can either disable the antivirus program temporarily, or add Ollama to your trust list.
• Firewall: Windows Firewall may prevent Ollama from downloading models, if you encounter problems with downloading, you can check your firewall settings to allow Ollama networking.