Cloud Studio 部署gemma3，映射本地端口到域名，进行外部访问

本文链接：https://blog.youkuaiyun.com/qq_32205577/article/details/146248819

Cloud Studio（云端 IDE）

选一个（试用 free）高性能workspce，约么16G显存

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 525.105.17   Driver Version: 525.105.17   CUDA Version: 12.0     |
|-------------------------------+----------------------+----------------------+
| GPU  Name        Persistence-M| Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp  Perf  Pwr:Usage/Cap|         Memory-Usage | GPU-Util  Compute M. |
|                               |                      |               MIG M. |
|===============================+======================+======================|
|   0  Tesla T4            On   | 00000000:00:09.0 Off |                    0 |
| N/A   48C    P8    15W /  70W |      5MiB / 15360MiB |      0%      Default |
|                               |                      |                  N/A |
+-------------------------------+----------------------+----------------------+
                                                                               
+-----------------------------------------------------------------------------+
| Processes:                                                                  |
|  GPU   GI   CI        PID   Type   Process name                  GPU Memory |
|        ID   ID                                                   Usage      |
|=============================================================================|
|  No running processes found                                                 |
+-----------------------------------------------------------------------------+

部署gemma3

16G的显存，选了12b的模型 ollama install，要是显存够用，模型size往上堆就完了
ollama
cloud studio workspace内置ollama服务，下载

ollama run gemma3:12b

ollama

如遇到ollama 错误

writing manifest 
success 
Error: llama runner process has terminated: this model is not supported by your version of Ollama. You may need to upgrade

需要升级内置的ollama版本，到v0.6.0

# 停止当前运行的 Ollama 服务
sudo systemctl stop ollama
# 下载最新安装脚本并重装
curl -fsSL https://ollama.ai/install.sh | sh
# 重启服务
sudo systemctl start ollama

查看ollama版本和model（需要重启）
ollama

绑定域名+端口映射

腾讯云CloudStudio：喂饭教程：10分钟部署一个DeepSeek，支持网页调用！关键还免费！

总结

这一套方案是将cloud studio的高性能工作空间做成了ollama service，并将本地ollama的端口映射到了公网域名，可以外部访问。