You guessed right, Flux can't do sweaty armpits either.
v5.5 is v5 with block analysis, and v6 is a new trained version, read version description. I recommend mixing both.
For one version alone, recommend weight 0.93
try from 0.8 to 1.2
Negatives (DynamicThresholding or Skimmed) can sometimes help, but not needed. You can download my images and drop on Comfy, the workflow has both options and a High-res upscale.
Trigger Keyword:
v7.5: sweat-damp on the armpit area, sweat-damp on the belly area, sweat-damp on the chest areaSupporting prompts:
Shirt with sweat, wet, HD high professional quality photo, best qualityNegative supporting prompts:
dry shirt, worst quality, jpeg artifacts, blurryFlux don't bleed insanely with the word "wet" as did SD15 and SDXL so it's safe to use it.
Trained with Ostris trainer.
old v1 description:
Very initial model, looks undertrained, but it's giving great results. My first experiment with Flux. This is the first time I'm not going to write 4 pages of info. You guys can chek the SDXL and SD15 version if you want. It's the same dataset.
I'll probably update this with new versions in the future.
I recommend weight 1.12
try from 0.8 to 1.5
Trained with Ostris trainer.
Description
Suggested strength: 0.93
This is a new trained version with kohya. A very low LR and trained on dev2pro. Also no regularization (and I think reg would help here) with block analysis remerged with my own tool with preset8: 0,0,0,0,0,0,0,0,0.25,0.25,0.25,0.25,0.25,0.5,0.25,0.5,0.5,0.25,0,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,0.25,0.25,0.25,0.25,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0
I think this performs better than v5.5, but i'ts not perfect. I think I overdid it on the low rank, and using this v6 and v5.5 together would be the better option.
Furthermore, I did not try to merge because merging never ends up any better than using both.
epoch 34 of 36, Prodigy, 1024x1024, steps ??60 of 6840,