wah. 21.
Suggestive at times
| Site (WIP) | http://pawgge.rs |
| https://twitter.com/zenithO_o | |
| Pronouns | He/Him |
| Shiny ✨🔞 | https://rubber.social/@zenith |
It has a total of 228 possible tags, all based off of e621's tagging scheme. It's categories are Rating, Species, and General tags.
It is also capable of tagging things like gender identity, body parts, clothing, and actions taking place within the image!
Hey everyone! I got a new project to share. I made a tool that can tag furry artwork!
Show me below the results you get! :3
Here is the architecture of the model. I start with a pre-trained CNN classifier as the base model, and then have multiple prediction heads for each category.
Every parameter is finetuned during training, meaning that the final output layer of the CNN model (the KerasLayer with an output of 1000) serves as a feature vector specifically for furry artwork.
Each prediction head also has a small fully connected Dense layer for category specific labels.
I did a thing! :D
Currently working on a model to label furry artwork, based off the e621 tagging system. It labels 6 different categories: Identity, Species, Body, Clothing, Action, and Rating.
These results are after only one epoch of training on my dataset and was the preliminary test run to see if this would even work. It does 👀👀👀
I had two completely different elevator experiences yesterday while holding my fursuit head.
My apartment’s elevator: *piss drunk students trying to hold back laughter from seeing a furry*
My friend’s apartment’s elevator: “YOOO IS THAT A FURRY MASK??? LEMME TRY IT ON”
Was funny to see the opposite ends of the spectrum XD
I've been working on a visualization for tag relationships on e621! This one in particular shows how general and species tags all relate to one another.
Here's a video explaining it and how the connections and colors work: