This LoRA was trained in about 4300 cartoon panels of Groo the Wanderer comics, using BLIP as captions. It acts differently than the first version, not necessarily better

There is no specific trigger word, but these are the most meaningful words in its training:

Count Name

----- ----

5786 man

4900 cartoon

1724 woman

1622 talking

1243 dog

1064 sword

1034 holding

858 bubble

807 head

782 hand

733 people

639 comic

577 hat

571 says

460 group

429 is

426 speech

375 thought

373 background

350 beard