ModuleNotFoundError: No module named 'diffusers.pipelines.pipeline_utils'

[swayam1998]

After following the fine-tuning/dreambooth/stablediffusion_dreambooth_finetuning.en.ipynb notebook steps I get this error.

PYTHONPATH=/opt/ml/code:/opt/conda/bin:/opt/conda/lib/python39.zip:/opt/conda/lib/python3.9:/opt/conda/lib/python3.9/lib-dynload:/opt/conda/lib/python3.9/site-packages:/opt/ml/code/repositories/xformers
Invoking script with the following command:
/opt/conda/bin/python train_dreambooth.py --attention xformers --class_data_dir /opt/ml/input/data/class_images/ --class_prompt photo\ of\ a\ man --gradient_accumulation_steps 1 --gradient_checkpointing True --instance_data_dir /opt/ml/input/data/images/ --instance_prompt photo\ of\ zwx\ man --learning_rate 2e-06 --lr_scheduler constant --lr_warmup_steps 0 --max_train_steps 300 --mixed_precision fp16 --model_name aws-trained-dreambooth-model --models_path /opt/ml/model/ --not_cache_latents True --num_class_images 50 --pretrained_model_name_or_path SG161222/Realistic_Vision_V1.4 --prior_loss_weight 0.5 --resolution 512 --sample_batch_size 1 --save_steps 300 --save_use_epochs False --train_batch_size 1 --train_text_encoder False --use_8bit_adam False --use_ema True --with_prior_preservation True
2023-03-28 08:26:02,862 sagemaker-training-toolkit INFO Exceptions not imported for SageMaker Debugger as it is not installed.
2023-03-28 08:26:02,862 sagemaker-training-toolkit INFO Exceptions not imported for SageMaker TF as Tensorflow is not installed.
Traceback (most recent call last):
File "/opt/ml/code/train_dreambooth.py", line 30, in
from extensions.sd_dreambooth_extension.dreambooth import xattention
File "/opt/ml/code/extensions/sd_dreambooth_extension/dreambooth/xattention.py", line 10, in
from diffusers.pipelines.pipeline_utils import LOADABLE_CLASSES
ModuleNotFoundError: No module named 'diffusers.pipelines.pipeline_utils'
2023-03-28 08:26:04,999 sagemaker-training-toolkit ERROR Reporting training FAILURE
2023-03-28 08:26:04,999 sagemaker-training-toolkit ERROR ExecuteUserScriptError:
ExitCode 1
ErrorMessage ""
Command "/opt/conda/bin/python train_dreambooth.py --attention xformers --class_data_dir /opt/ml/input/data/class_images/ --class_prompt photo\ of\ a\ man --gradient_accumulation_steps 1 --gradient_checkpointing True --instance_data_dir /opt/ml/input/data/images/ --instance_prompt photo\ of\ zwx\ man --learning_rate 2e-06 --lr_scheduler constant --lr_warmup_steps 0 --max_train_steps 300 --mixed_precision fp16 --model_name aws-trained-dreambooth-model --models_path /opt/ml/model/ --not_cache_latents True --num_class_images 50 --pretrained_model_name_or_path SG161222/Realistic_Vision_V1.4 --prior_loss_weight 0.5 --resolution 512 --sample_batch_size 1 --save_steps 300 --save_use_epochs False --train_batch_size 1 --train_text_encoder False --use_8bit_adam False --use_ema True --with_prior_preservation True"
2023-03-28 08:26:04,999 sagemaker-training-toolkit ERROR Encountered exit_code 1

I think the diffuser version in the public docker image is causing the issue. I do see some commits by @stevensu1977 to solve this issue but they are still occurring. Am I going wrong somewhere?