large(maybe) tests show:rv1.2 performs well with easynegativerv1.4 performs well with no easynegativeand the only difference between them is training steps.<hr /><h3 id="heading-81">100% game textures in dataset.</h3>Highly recommended:furry, rabbit ears, rabbit nose, :-&lt;, yuran_face, yuran_body, yuran_tail,yuran_hairstyle in prompt andsimple background, black background in negative promptNegative prompt above is highly recommended because of the feature of the dataset.<hr />The fact is, prompts depends largely on your base model.After I run all the posts, I changed the Lora file name for easy understanding. Better pay attention when copy the generation data.

Files with SHA256: 5797f59d279f1fb52c136f1e82c266b18941580b5b9e60e0dc04bf5118c4d8fb