3 contributors

History: 6 commits

PseudoTerminal X

Addressing a training quick from earlier SimpleTuner versions

7e84121 verified 6 months ago

scheduler
Update scheduler/scheduler_config.json 11 months ago
text_encoder
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
text_encoder_2
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
tokenizer
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
tokenizer_2
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
unet
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
unet_ema
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
vae
v1: 38800 + 21800 + 15400 steps of mixed LAION/MJ dataset and offset noise with input perturbation on a probability of 25%, 10% caption dropout, cosine LR 4e-7 to 8e-7 every 3200 steps, no min-snr loss augmentation about 1 year ago
.gitattributes

1.52 kB

initial commit about 1 year ago
README.md

4.29 kB

Update README.md about 1 year ago
model_index.json

620 Bytes

Addressing a training quick from earlier SimpleTuner versions 6 months ago