Quickly experience DeepSeek, supporting models sized 1.5b/7b/8b/14b/32b/70b, with no need to wait for model downloads.

1. Fork the repository to your organization.
2. Select your preferred branch.
3. Click on Workspace.
4. After approximately 5-10 seconds, in the remote development command line, enter the following command to experience DeepSeek:

ollama run $ds

If you only need to have conversations with DeepSeek:

• Click on the search bar in the upper right corner of the page and type your question directly.
• Alternatively, press the / key for an enhanced experience.

To run DeepSeek locally within your CVM (taking 14b as an example), follow these steps:

docker run --rm -it docker.cnb.cool/examples/ecosystem/deepseek/14b:latest
ollama serve &
ollama run deepseek-r1:14b

Note: The domain docker.cnb.cool is globally accelerated within Tencent Cloud's intranet, ensuring zero traffic fees.

To develop AI applications based on DeepSeek, you can deploy using the following methods:

• Develop and Continuously Deploy AI Applications Based on DeepSeek to HAI
• Deploy DeepSeek to a Specified VPC Using Cloud Applications