Support token counting for newer models before sending request to OpenAI

iterativecloud / jupyter-notebook-chatcompletion

Jupyter Notebook ChatCompletion is VSCode extension that brings the power of OpenAI's ChatCompletion API to your Jupyter Notebooks!

https://www.iterative.cloud/

MIT License

6 stars 2 forks source link

Support token counting for newer models before sending request to OpenAI #11

Closed AlexandreDeRiemaecker closed 4 months ago

AlexandreDeRiemaecker commented 4 months ago

When a user selects some of the latest OpenAI models (given that they have been given access to them by OpenAI), the following warning will appear "Warning: Error while counting tokens - skipping token limit checks".

This is likely to be resolved with an update of the tiktoken dependency - or, in the worst case, a fallback to the next best approximation for counting tokens.

AlexandreDeRiemaecker commented 4 months ago

Resolve with #14.