⏩ 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
Describe the bug
When configuring an Azure OpenAI deployment, Continue uses the model name to check max tokens. However, it uses the OpenAI model names which differ from those used by Azure OpenAI (gpt-4, gpt-35-turbo, etc). As you can expect, hilarity and KeyError ensue. The only way to get past this is to check Continue's source code and see what model names it would expect.
To Reproduce
Configure Continue as per the docs, and use an Azure OpenAI model name such as gpt-35-turbo-16k.
Expected behavior
The extension doesn't crash complaining about KeyError.
Describe the bug When configuring an Azure OpenAI deployment, Continue uses the model name to check max tokens. However, it uses the OpenAI model names which differ from those used by Azure OpenAI (
gpt-4
,gpt-35-turbo
, etc). As you can expect, hilarity andKeyError
ensue. The only way to get past this is to check Continue's source code and see what model names it would expect.To Reproduce Configure Continue as per the docs, and use an Azure OpenAI model name such as
gpt-35-turbo-16k
.Expected behavior The extension doesn't crash complaining about KeyError.