Type | |
Stats | 1,264 |
Reviews | (283) |
Published | Sep 4, 2024 |
Base Model | |
Training | Steps: 307,200 Epochs: 80 |
Usage Tips | Clip Skip: 1 |
Hash | AutoV2 7285BD92B3 |
Hi there and welcome to a merge of 2 finetunes to get some 2.5D looking checkpoint, ended up liking the current result a lot, so ima release it as is, shiny skin, oily skin, and oiled skin give the look I like from this ckpt, give it a try!
You can gen with or without Score tags, and up to cfg 8!
Emphasis mode No Norm is suggested for SDXL models.
Huge thanks to @Anzhc with his help in training hyperparameters, and @renmei for motivating me to do this
If you want bigger girls go and use mnemosynekotorin's Venus Body lora
Positive prompt : score_9, score_8_up
Helper prompts : realistic, 3d
Negative prompt : source_pony, source_furry, source_cartoon, signature, watermark, username, monochrome, censored, mosaic censorship, censorship
Personal schizo Negative prompt: worst quality, low quality, sketch, normal quality, jpeg artifacts, depth of field, blurry, bloom, messy drawing, amateur drawing, fewer digits, extra digits, source_pony, source_furry, source_cartoon, signature, watermark, username, monochrome
Sampler : Euler A AYS, Any Align Your Steps even GITS.
CFG : 4 to 7 (can be pushed up to 9 and it will work on a lot of prompts
If you want to support me, consider liking the model, and/or tipping buzz/donating to my Ko-fi, as this was finetuned on 4xA100 for around 307200 steps.
Relevant training information:
Trained with at least 7k images.
Total of 307,200 real steps after batching across 80 epochs.
Effective batch size was 40 x 4 GPUs x 4 Gradient Accumulation
Learning rate was 3e-6, TE1 was 1e-7, TE2 was 8e-8