[launch_inpaint.sh] AttributeError: 'LatentsDataset' object has no attribute 'class_images_path'

[ZeroCool22]

Describe the bug

My launch_inpaint.sh

export LD_LIBRARY_PATH=/usr/lib/wsl/lib:$LD_LIBRARY_PATH export MODEL_NAME="runwayml/stable-diffusion-inpainting" export INSTANCE_DIR="training" export OUTPUT_DIR="my_model"

accelerate launch train_inpainting_dreambooth.py \ --pretrained_vae_name_or_path="stabilityai/sd-vae-ft-mse" \ --pretrained_model_name_or_path=$MODEL_NAME \ --instance_data_dir=$INSTANCE_DIR \ --output_dir=$OUTPUT_DIR \ --instance_prompt="Greyalieninpa" \ --resolution=512 \ --train_batch_size=1 \ --learning_rate=1e-6 \ --lr_scheduler="constant" \ --lr_warmup_steps=0 \ --gradient_accumulation_steps=4 --gradient_checkpointing \ --use_8bit_adam \ --save_interval=500 \ --max_train_steps=5000

Reproduction

No response

Logs

No response

System Info

• diffusers version: 0.9.0
• Platform: Linux-5.10.102.1-microsoft-standard-WSL2-x86_64-with-glibc2.31
• Python version: 3.9.13
• PyTorch version (GPU?): 1.12.1+cu116 (True)
• Huggingface_hub version: 0.10.0
• Transformers version: 4.24.0
• Using GPU in script?: <1080 TI>
• Using distributed or parallel set-up in script?:

[InB4DevOps]

add --not_cache_latents to your launch_inpaint.sh and it should work

[vivek-hirer-ai]

--not_cache_latents leads to an out of memory error even on a 24gb system. Was there any workaround found to this? Running into the same issue with train_inpainting_dreambooth.py as the one mentioned i.e. 'LatentsDataset' object has no attribute 'class_images_path'