controlNet-size=512-batchsize=6-epoch=1

accelerate launch train_controlnet.py \ --pretrained_model_name_or_path=$MODEL_DIR \ --output_dir=$OUTPUT_DIR \ --dataset_name=$DATA_DIR \ --resolution=512 \ --learning_rate=1e-5 \ --image_column=image \ --caption_column=caption \ --conditioning_image_column=grayscale_image \ --train_batch_size=6 \ --gradient_accumulation_steps=4 \ --num_train_epochs=1 \ --enable_xformers_memory_efficient_attention \ --checkpointing_steps=1000 \ --validation_steps=500 \

Zixian Yan

Created on October 19|Last edited on October 19

Comment

﻿
Section 1﻿
Run: happy-meadow-91
﻿
﻿
﻿
Run: happy-meadow-91
﻿
﻿

Add a comment