LoRa is a neural network model based on SD 1.5. She generates furry art. The model was trained on a dataset consisting of more than 600 images.
The activation word for this model is 'qttips'. This word is used to activate the image generation process.
In the dataset on which the model was trained, the most frequently occurring elements were: background, furry, hair, animal, ears, body, eyes, tail, simple and fur ).
The model went through 50 epochs of training with 10 steps per epoch. This is just the beginning of training, and in the future it is planned to expand the dataset and increase the number of epochs to improve the quality of the generated images.
This neural network model is a great choice for furry art lovers, and it continues to improve with each new era of training.
Here are some examples of version 1.0 working:
(This is not the best example, there are much better ones)
Future plans include further improvement of this model. This will happen by expanding the dataset and increasing the number of training epochs. In this way, the quality of the generated images will constantly improve.