系统变量，对系统的所有账户生效，您可以根据情况自己选。点击添加添加，输入变量名OLLAMA_MODELS，变量值输入您放置ollama模型的新地址，我这里放在了E:\ollama\models

  1. 拉取模型

ollama pull deepseek-r1:8b

  1. 运行模型

ollama run deepseek-r1:8b

🔍 Can I run DeepSeek-R1:8B via Ollama?

Currently, Ollama does not officially support the DeepSeek-R1 8B model unless it has been added to their
repository or community-supported list. You should check:

  • The Ollama models page for updates.
  • Community forums (e.g., GitHub issues) if users have shared custom Docker images/definitions.

🧪 How to run transformer models like DeepSeek-R1 via Ollama

If the model is supported, you can use this command:

1
ollama run deepseek-r1:8b

This will:

  • Download (if not installed) and load the 8B parameter version of DeepSeek-R1.
  • Start a local instance for interactive chat or API access.

🤔 What is DeepSeek-R1:8B?

DeepSeek-R1 is an advanced transformer language model developed by 深度求索（DeepSeek） . The “8B” refers to 8 billion parameters in the model. It’s designed for:

  • Natural language understanding and generation.
  • Code-related tasks (syntax, debugging).
  • Knowledge answering.
  • Creative writing.

📄 Example usage with Ollama CLI

1
2
3
4
5
6
7
8
9
10
11
12
# Start interactive chat (if supported)
ollama chat deepseek-r1:8b

# Or generate text via API/command line:
curl http://localhost:11434/api/generate \
  -X POST \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-r1:8b",
    "prompt": "Explain quantum computing in simple terms.",
    "stream": false
  }'

⚠️ If the model isn’t supported yet:

You can try running it via Docker (if you have technical access):

1
2
3
4
5
6
docker run -d --name deepseek-r1 \
  -p 11434:11434 \
  --rm \
  -v ollama_models:/models \
  llamafile/docker-ollama \
  predict DeepSeek-R1 /models/deepseek-r1.tar

But note: This may not be officially maintained.

🌟 What else can Ollama run?

Oll (a) supports models like:

  • Llama, Mistral, Mixtral
  • GPT models (via LMSYS)
  • Gemma, Phi, CodeLlama
  • And many others!

Check the current list on their website.

Let me know if you’d like help with a specific task using DeepSeek-R1 or another model! 😊