support Qwen/Qwen2.5-Coder-32B-Instruct model provided by siliconflow

[AnoyiX]

Description

support some models provided by siliconflow, like

• Qwen/Qwen2.5-Coder-32B-Instruct
• Qwen/Qwen2.5-Coder-7B-Instruct

Testing

custom config:

{
  "models": [
    {
      "title": "Qwen2.5 Coder",
      "provider": "openai",
      "model": "Qwen/Qwen2.5-Coder-32B-Instruct",
      "apiKey": "************************************",
      "apiBase": "https://api.siliconflow.cn/v1/"
    }
  ],
  "tabAutocompleteModel": {
    "title": "Qwen2.5 Coder",
    "provider": "openai",
    "model": "Pro/Qwen/Qwen2.5-Coder-7B-Instruct",
    "apiKey": "************************************",
    "apiBase": "https://api.siliconflow.cn/v1/"
  },
  ...
}

AI Edit Example:

Autocomplete Example:

[netlify]

✅ Deploy Preview for continuedev ready!

Name	Link
🔨 Latest commit	26e0017
🔍 Latest deploy log	https://app.netlify.com/sites/continuedev/deploys/6746b793dcf4c4000868435d
😎 Deploy Preview	https://deploy-preview-2974--continuedev.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[sestinj]

@AnoyiX Can we make a dedicated LLM class instead of putting the logic in OpenAI.ts? A good example would be in Deepseek.ts, where we subclass OpenAI and adjust the FIM logic

[AnoyiX]

It might be better to allow custom endpoint, there are so many providers, it’s hard to support them all, what do you think?

[sestinj]

I think that's fine as long as they have the exact same expected request body. I would just make it a function like getFimEndpoint so that it can be modified by subclasses, and then make a subclass of the endpoint. I'm not worried about having a large number of subclasses

[AnoyiX]

Sounds good. It's convenient for subclasses to implement getFimEndpoint, this way we can avoid duplicating a lot of code in *_streamFim

[AnoyiX]

you can review the latest code ~ 😊