Update README.md

README.md

@@ -4,8 +4,10 @@ license: creativeml-openrail-m
 
 # Fubuki (one punch man)
 Base model: https://huggingface.co/Linaqruf/anything-v3.0
+Used 'fast-DreamBooth' on Google Colab, 7600 steps, fp16, 640x640 images.
 
-Trained with 38 yusuke murata style (and similar) 640x640 images, hand-picked and hand-cropped, some edited to remove elements from the background. Not trained with nsfw images, can do nsfw but it's limited (see nsfw examples).
+Trained with 38 yusuke murata style (and similar) images, hand-picked and hand-cropped, some edited to remove elements from the background. Not trained with nsfw images, can do nsfw but it's limited (see nsfw examples).
+Text-encoder trained with 43% of steps.
 
 I recommend going with no more than 832x832 and not less than 512x512, stay near 640x640 or 704x704 for better results.