The latest version of PixelWave with better cohesion, detail and prompt following skills!
Trained on a lot of pictures captioned with ChatGPT 4 vision. Also added in the art used to train the PixelShade model. So it can do some cool artsy styles.
It also outputs great quality images without hires fix!
No need for the refiner model.
This model is not a mix of other models.