How do I add a fursuit to be recognized by the bot?

Machine Learning works from examples. It starts off with correctly-named examples, then it works out what features are relevant to tagging the photo with each name.

The most important need is a significant number of photos that are of the fursuit that is to be added.

By significant number, I suggest as a rule of thumb, at least 30. On the higher end, some of the fursuits have over 200 labelled photos, which also works.

The model is not directly crowdsourced, I upload the photos myself, after checking them. A couple options:

• You can upload them to me on Telegram. My user name is @RainRattie. To save the time of individually uploading each one, put them all in one ZIP file, and use the Send File function.
• Or, if you don't know about ZIP files, send them as photos to the bot. As mentioned in the other post, I do look at submissions. If I'm looking for a specific fursuit, I can grab the photos from the incoming files.

Then you have to tell me the name of the fursuit. If you messaged me the photos directly, you can tell me the name as you send it. If you sent the photos to the bot, also message me, but tell me what to look for in the incoming files.

Some suggestions:

• The only fursuit in the photos should be the one that you are adding.
• Backgrounds are ok; they can be busy, plain, light, dark, outdoors, indoors. In fact, please change up the backgrounds as much as possible.
• Change up the lighting. It doesn't mean you have to take photos that are super-dim or super-bright. Just that fursuits look quite different under sunlight, incandescent, and florescent lights.
• Change up angles and poses. For instance, if all your photos are straight-on of the face, and the model is trained on that, later when someone sends a photo of a side view, it's going to look totally different and going to be unrecognized.