This model removes belly buttons. No keyword required.How it was trained:<ul><li>Lora overtrained on an image</li><li>Second lora overtrained on an edit of that image (in this case the belly button was inpainted out)</li><li>Repeated the above a few times</li><li>Merged all the loras trained on edited images into checkpoint_a</li><li>Merged all the loras trained on unedited images into checkpoint_b</li><li>Extracted lora from checkpoint_a - checkpoint_b</li><li>Zeroed out non-required blocks in the lora using trial and error to reduce lora side effects on generated image</li></ul>Used the following guide:<a target="_blank" rel="ugc" href="https://civarchive.com/models/98334/difflorafaceshadowtweaker">https://civarchive.com/models/98334/difflorafaceshadowtweaker</a> <a target="_blank" rel="ugc" href="https://1lemon1lemon.netlify.app/2023-06-27_difflora/">https://1lemon1lemon.netlify.app/2023-06-27_difflora/</a>Used this for training:<a target="_blank" rel="ugc" href="https://github.com/bmaltais/kohya_ss">https://github.com/bmaltais/kohya_ss</a>Used the following command for overtraining loras (paths replaced with ...):<pre><code>accelerate launch --num_cpu_threads_per_process=2 "train_network.py" --enable_bucket --pretrained_model_name_or_path="..." --train_data_dir="..." --resolution="768,768" --output_dir="..." --logging_dir="..." --network_alpha="16" --save_model_as=safetensors --network_module=networks.lora --unet_lr=0.001 --network_train_unet_only --network_dim=32 --output_name="01a" --lr_scheduler_num_cycles="1" --learning_rate="0.001" --lr_scheduler="constant" --train_batch_size="1" --max_train_steps="400" --save_every_n_epochs="1" --mixed_precision="fp16" --save_precision="fp16" --cache_latents --optimizer_type="AdamW8bit" --max_data_loader_n_workers="0" --bucket_reso_steps=64 --xformers --bucket_no_upscale</code></pre>Used this A1111 extension for merging:<a target="_blank" rel="ugc" href="https://github.com/hako-mikan/sd-webui-supermerger">https://github.com/hako-mikan/sd-webui-supermerger</a>

Files with SHA256: fd1035c4dc0ea210b931286aba2930b24b0bee7b0006bcac6d4221b0571121ee