search

continuedev / continue

⏩ Continue is the leading open-source AI code assistant. You can connect any models and any context to build custom autocomplete and chat experiences inside VS Code and JetBrains
https://docs.continue.dev/
Apache License 2.0
19.25k stars 1.66k forks source link

Fix config schema for tab autocomplete models #2839

Closed amytimed closed 6 days ago

amytimed commented 1 week ago

Old schema used a key which isn't used by Continue, and thus you couldn't specify tab autocomplete model by following schema. This updates it to how it is in the the real

netlify[bot] commented 1 week ago

Deploy Preview for continuedev failed. Why did it fail? →

Name Link
Latest commit 6085199d3cee71f8dabad3820eb53f3808a38f49
Latest deploy log https://app.netlify.com/sites/continuedev/deploys/672d1c38c667d800081d220f
RomneyDa commented 1 week ago

Hi, @amytimed, thanks for the PR!

Mentioned in discord

Hi @Amy, in config.json it uses tabAutocompleteModel, which can be set to one model or an array of models. Then it converts from intermediate to final config here, and the s is added for the final type https://github.com/continuedev/continue/blob/dev/core/config/load.ts Adding the s fixed it because now it's not finding your tabAutocomplete model so it's defaulting to starcoder2:3b, which is not causing the memory issue It's a bit confusing but the root issue seems to be that deepseek-coder:1.3b-base is somehow requesting way too much memory, or perhaps we're not sending some param that deepseekcoder needs for memory management. Might be related to this https://github.com/continuedev/continue/pull/2793

sestinj commented 6 days ago

As mentioned above, "tabAutocompleteModels" actually isn't intended to be supported, so going to close. Thank you though for taking the time to open a PR!