Trigger words
To trigger the LoRA use the prompt in style of DSYL
Training
This adapter has been trained using the autotrain-advanced dreambooth colab-notebook
with the dataset stored in ./images/training
.
Training time was ~14 minutes on a T4 instance.
Config
The following config has been used
project_name = 'disco-elysium'
model_name = 'stable-diffusion-v1-5/stable-diffusion-v1-5'
prompt = 'in the style of DSYL'
learning_rate = 1e-4
num_steps = 500
batch_size = 1
gradient_accumulation = 4
resolution = 512
use_8bit_adam = False
use_xformers = False
mixed_precision = "fp16"
train_text_encoder = True
disable_gradient_checkpointing = False