TREAT-Classic / model

Commit History

Switched back to base for 20% increased efficiency
75877cb
Running

Kuberwastaken commited on

Experimenting with t5 large model
6538671

Kuberwastaken commited on

Reverting back to old working model.
e84c507

Kuberwastaken commited on

Removed Dual start and stop APIs in gradio app
4a39e77

Kuberwastaken commited on

Added a manual port to sort runtime errors
e2afe19

Kuberwastaken commited on

Added actual trigger categories instead of dummy ones
07f4e87

Kuberwastaken commited on

Worse Prompts are broken too now
4ffd7e0

Kuberwastaken commited on

worse prompts = better results
bf69cf6

Kuberwastaken commited on

something broke, trying model without the loading bar.
8ea6312

Kuberwastaken commited on

Improved model and functioning loading bar
474b075

Kuberwastaken commited on

one step forward, two steps back
0b22cf5

Kuberwastaken commited on

Somehow didn't break the model, so pushing it further
fdb321f

Kuberwastaken commited on

This will probably break the model but ah well
6509cc1

Kuberwastaken commited on

Making model more deterministic
0db78ec

Kuberwastaken commited on

I have no clue why this works, but this works
c689a93

Kuberwastaken commited on

Made model actually work again
00636d1

Kuberwastaken commited on

Minor UI updates, made model better
da1a8ef

Kuberwastaken commited on