About this version (qwen v2)
A brushup of the dataset, with better result than V1 for qwen. I will use the same dataset again for zturbo when the base model is released.
Trigger word: PENISLORA
What can this lora do?
This lora can add erect penises to both men or women viewed from the front/side. Other angles such as POV may have a backwards penis head.
Other things it can now do:
Side view of the penis
Cumming / Cumshots
Blowjobs (its captioned for the words "blowjob" and "deepthroat" )
What can't it do?
No penetration in the training data. Also nothing from POV angle, though there is a few images from above and 1 POV video in the training data.
Sometimes blowjobs with cumming have the penis slip out the closed mouth.
Recommended Settings
It works pretty good with the new lightning dyno high model. I'll link to it in my example workflow. I like to use dyno high model (no lightning lora), then for low I use the lightning v2 lora on the regular 2.2 low base model.
Dataset
84 images at 512x resolution
43 videos at 256x resolution
(I let DP pick the aspect ratio automatically)
This is the same exact dataset as the 2.2 5B model. I made no changes.
Training
I used the default diffusion pipe settings.
[optimizer]
type = 'adamw_optimi'
lr = 2e-5
betas = [0.9, 0.99]
weight_decay = 0.01
eps = 1e-8
I was baffled why it was taking so long to train the high until I realized after over 60 hours of training that I had put my videos in the images directory which resulted in the high being trained ONLY only on videos and twice (once with a very high resolution). Once I fixed this, I went back and trained from 11K steps up to around 13K with the images in the training data. The high model was fine without to be honest.
For the low, I trained it properly with videos and images the whole way, around 6K steps in I upped the image resolution from 512 to 1024 actually and didn't get an OOM (it fit around 24GB exactly). I trained it to around 10.5K steps. Also I trained the low on the full timestep range (0 to 1 instead of 0 to 0.85) from some advice, it may switch better over from high to low on the speed up lora with low steps.
I think I might do another version with more angles such as POV and from the behind to make this work for any situation. In that case I don't think it needs 10K steps per training session, epochs around 5K steps looked fine.
The results
I think it was a combination of improved captioning and 2.2 base model being better. But this lora turned out really well.
Description
Notes about 1.3B version which is whats new here:
Warning: I haven't done much in terms of testing, this may be overtrained. Try the epoch99 and compare to 109, 99 had good results and I haven't had time to do comparisons or do many generations at 109epochs. I'll do some tests later and if its better then I will revert to 100 epoch
Trigger: PENISLORA
This is a 1.3B T2V version of my penis lora. It does for the most part everything the 14B version does(though a bit more poorly).
Trained to around 109 epochs and 23,900 steps (109 hours, around 4.5 days straight of training locally on my 3090). Unless I add new training data, I think we are starting to get into over trained territory. In the coming days, I will work on adding more data and retraining the 14B model lora and if it works out then I can update this 1.3B with a v0.8. Check the example generations for example prompts or try from those listed in 14B v0.7 lora.