What can I help with?
I can’t keep the server running 24/7 because of the cost. If it’s asleep, your first reply may take several minutes while it starts.
About this chat
This project runs on a tight budget and uses CPU inference. If many people are chatting at once, you may have to wait in a queue.
The model is an experimental, small model built to beat GPT-2. It can make things up confidently, so please check important information. Thanks for your patience while this project grows.
Experimental small model · Responses may be unreliable