The rate limiter seems to be oversensitive, at least to as what counts as a message

#218
by BenSmithh - opened

The rate limiter seems to be oversensitive, at least to as what counts as a message;

I have been using the llama 2 model on huggingchat recently and have noticed doing Web Searches always seems to fill my rate limit, sometimes only getting one message in before hitting the limit.

I'd recommend changing what classes towards the rate limit, or limiting the amount of requests the web search makes.

Sign up or log in to comment