完了它一晚上，太好玩了。真的不敢相信 32b parameter model 能这么快这么好，一个 RTX 3090 用起来完全没问题
大家快试试
Instructions are for Windows machines
Prerequisites
Download Ollama Download Ollama on macOS
Download Docker Get Docker | Docker Docs
Run Ollama (it will run in the background)
Open a command prompt or PowerShell window and type one of these:
1.5B version (smallest):
ollama run deepseek-r1:1.5b
8B version:
ollama run deepseek-r1:8b
14B version:
ollama run deepseek-r1:14b
32B version:
ollama run deepseek-r1:32b
70B version (biggest/smartest):
ollama run deepseek-r1:70b
EDIT：有人做出来没有审查的版本，需要使用可以用以下 command
32B version:
ollama run huihui_ai/deepseek-r1-abliterated:32b
or
ollama run huihui_ai/deepseek-r1-abliterated:32b-qwen-distill
70B version:
ollama run huihui_ai/deepseek-r1-abliterated:70b
more info here
huihui_ai/deepseek-r1-abliterated
Pick your parameter size based on your graphics card (32b model 对我的 RTX 3090 非常友好，22/24 GB VRAM used):
Then type
docker run -d -p 3000:8080 --add-host=host.docker.internal:host-gateway -v open-webui:/app/backend/data --name open-webui --restart always ghcr.io/open-webui/open-webui:main
Then go to http://localhost:3000
Use any name and email; it’s stored locally
然后就好了
试水了以下LMStudio，下载完直接运行
