Open-Assistant
Open-Assistant copied to clipboard
Create new page to query a live model
This new page should query the path made in #425 with a prompt written by the user and then display the results. The user should be able to request a response from one or more models in a drop down. Those model IDs can be hard coded for now.
It should also make it easy to reset the view and start over.
The inference server now exists. Once there's a docker setup, we should use the interactive text-client example to prototype live model interaction.
This should be blocked by a feature flag since it won't be ready for real usage for a while and we need to frame how and when users should use it.
This seems to have been resolved by #1915
Actually - we can leave this open for now, there are elements of this ticket which are not complete e.g. the dropdown to select model
I will close this now, even though we don't have the drop down, I don't think we will need it, and we can always create a new issue.