https://github.com/n3ddu8/deepseek-pi
Run Deepseek R1 on a Raspberry Pi
https://github.com/n3ddu8/deepseek-pi
ai ansible ansible-playbook chatbot deepseek deepseek-r1 docker docker-compose raspberry-pi
Last synced: about 1 year ago
JSON representation
Run Deepseek R1 on a Raspberry Pi
- Host: GitHub
- URL: https://github.com/n3ddu8/deepseek-pi
- Owner: n3ddu8
- License: mit
- Created: 2025-02-01T14:35:53.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2025-02-02T11:32:38.000Z (over 1 year ago)
- Last Synced: 2025-03-27T22:25:17.301Z (over 1 year ago)
- Topics: ai, ansible, ansible-playbook, chatbot, deepseek, deepseek-r1, docker, docker-compose, raspberry-pi
- Homepage:
- Size: 5.86 KB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Deepseek Pi
Using Ansible and Docker Compose, run Ollama with Open WebUI to host and interact with models such as Deepseek R1. This is just a bit of fun, and isn't really usable in real-world scenarios. For comparison, the Deepseek R1 that at the time of writing is supposedly beating ChatGPT runs with 671 billion parameters, whereas the the only version that works on a standard Pi uses 1.5 billion parameters.
## Setup
### Inventory File
Configure `inventory.yaml` with the details of the target machine.
### SSH
Run the following, following the onscreen instructions where necerssary:
```shell
ssh-keygen
ssh-copy-id @
ansible -i inventory.yaml all -m ping
```
This should return a `pong`.
Now run:
```shell
ansible-playbook -i inventory.yaml playbook.yaml
```
## Usage
Navigate to `http://:3000` and create an admin login. Now navigate to `http://:3000/admin/settings/` and click on `models` and `manage models` (looks like a small download link). In the `Pull a model from Ollama.com` box enter `deepseek-r1:1.5b` or whatever model you want to run.