3 ms·
Well done! I wonder if it's worth it to also hide a "select model" option somewhere, since we're putting in our own api keys i'd be interested in trying out oth
by ICodeSometimes 4y ago
Well done! I wonder if it's worth it to also hide a "select model" option somewhere, since we're putting in our own api keys i'd be interested in trying out other models even though they arent "chat" models.
Also would be great to see which messages are included in the computation, since chat gpt limits total characters per request to 4K? so inevitably some past messages will be rendered useless.
Overall well done!
- trungdq88 4y agoThank you for the suggestion! I'll add an option to select the model soon. As per your question, currently all the previous messages of the same chat will be included in the request. If you hit the 4k token limit, there will be an error message. I'll look into improving this too.