V2
I removed watermarks from the training dataset and also increased the diversity of shots and included more examples with clothing.
V2
I removed watermarks from the training dataset and also increased the diversity of shots and included more examples with clothing.