-
Notifications
You must be signed in to change notification settings - Fork 0
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Some simple suggestions #1
Comments
Hi! Thank you for the suggestions.
Hmm, but then too many tags might appear and might confuse users.
Yeah, that would be good. @onTDB needs to check this (since the API doesn't have character confidence yet).
This is because we are using a modified version of deepdanbooru with some changes in its neural network. |
Ten tags are too few, and there are generally not many tags that exceed the confidence threshold of 0.5. Most importantly, the 10 tags with the highest confidence levels tend to be too broad, like |
Hello, VoidxHoshi
That's a pretty good suggestion |
So I was thinking of the first enhancement and seems like we can first show only 10 tags and then if there're more tags users can click the Check the development site and try it |
First of all, thank you for your work, but I do have some small suggestions.
First, there are too few tags, only 10, it would be better to show all tags that exceed a confidence threshold of 0.5.
Second, for the character judgement, it would be better to add a confidence threshold and only show the character tags if they exceed the threshold (preferably also 0.5). There is not necessarily only one character, and a character may have multiple tags (corresponding to different forms, for example,
taihou_(azur_lane)
andtaihou_(phoenix's_spring_song)_(azur_lane)
) indanbooru
.Third, the calculation of rating is different from the original
deepdanbooru
, and the result is a bit strange.The text was updated successfully, but these errors were encountered: