add custom endpoint for AI suggestion

[yann300]

create a script in Remix:
like

(async () => {
  try {
    remix.call('copilot-suggestion' as any, 'useRemoteService', 'http://replit.remixcompletion.pro/infer')
    console.log('done')
  } catch (e) {
    console.log(e.message)
  }
})()

and run it to use a remote endpoint instead of the in-browser model.

[netlify]

✅ Deploy Preview for remixproject ready!

Name	Link
🔨 Latest commit	c588e5f
🔍 Latest deploy log	https://app.netlify.com/sites/remixproject/deploys/6596bf0840335e0008571607
😎 Deploy Preview	https://deploy-preview-4319--remixproject.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.

[STetsing]

@yann300 where am I supposed to execute that JS?

[yann300]

@STetsing you have to run it from Remix.

[STetsing]

The results is not displayed in the editor, Increasing the request timeout (if existent) could help

[STetsing]

• The replit endpoints works well
• the starcoder endpoint ends up with a Failed to load resource: net::ERR_FAILED starcoder.remixcompletion.pro/infer

[yann300]

• The replit endpoints works well
• the starcoder endpoint ends up with a Failed to load resource: net::ERR_FAILED starcoder.remixcompletion.pro/infer

the second issue looks like a server issue, is this the case?

[STetsing]

• The replit endpoints works well
• the starcoder endpoint ends up with a Failed to load resource: net::ERR_FAILED starcoder.remixcompletion.pro/infer

the second issue looks like a server issue, is this the case?

no, the server returns a payload indeed, what is the request max timeout? the min inference time for starcoder ti max_new_word=5 is 6-5 seconds (mean value).

[yann300]

what is the request max timeout

I think this can be configured by passing some options to https://github.com/ethereum/remix-project/pull/4319/files#diff-456892ab56ea0fe769bc1bdb2c0ece2855f1ce80583e2309939106e9cf727f0eR68 but I don't know the default value..

[STetsing]

The endpoints works well after reducing the max number of tokens (especially for the starcoder model). The timeout value must be increased for large generation requests

[STetsing]

Working as Is