Everyone's favorite questionable redesign.
Don't expect miracles from this one. The dataset is a bit limited as there's not much out there yet for this design. I recommend using it at a weight of 0.8 or lower or you'll start to get artifacting. Also the headpiece is kinda stuck on, you can try adding the tags to the negatives to it but it might not always work.
I don't really have a strong opinion on this design TBH. I was just bored and had an hour or two spare so I quickly made this. Hence the limited sample pics. Still, hope I made the 3 people out there who wanna goon to this design happy.
Description
Trained on 26 images. Epoch 8 was used.
{
"engine": "kohya",
"unetLR": 1,
"clipSkip": 2,
"loraType": "lora",
"keepTokens": 1,
"networkDim": 16,
"numRepeats": 8,
"resolution": 1024,
"lrScheduler": "cosine",
"minSnrGamma": 5,
"noiseOffset": 0.1,
"targetSteps": 624,
"enableBucket": true,
"networkAlpha": 8,
"optimizerType": "Prodigy",
"textEncoderLR": 1,
"maxTrainEpochs": 12,
"shuffleCaption": true,
"trainBatchSize": 4,
"flipAugmentation": false,
"lrSchedulerNumCycles": 3
}
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.