Changes:
Retrained on the smaller subset of original images after dropping images with curly hair. This results in better hair control now.
Merged second training with LoRa of the first version to obtain results than either approach overall.
Usage:
TiffanyTaylor <lora:TiffanyTaylor_M2:1.0>
Weights can be 0.9 to 1.2 depending on shot size (close-up, medium body or full body).
Currently up to medium body shows good results.
Changes:
Retrained on the smaller subset of original images after dropping images with curly hair. This results in better hair control now.
Merged second training with LoRa of the first version to obtain results than either approach overall.
Usage:
TiffanyTaylor <lora:TiffanyTaylor_M2:1.0>
Weights can be 0.9 to 1.2 depending on shot size (close-up, medium body or full body).
Currently up to medium body shows good results.