A collection of Python scripts for calling the REST API for the Dreambooth extension to the Automatic1111 Stable Diffusion WebUI

Overview

This is a collection of Python scripts for calling the REST API of the Dreambooth extension for the AUTOMATIC1111 Stable Diffusion Web UI.

After extensive testing, I have determined that the v1.2.1 release of the Web UI, and the 1.0.14 (unrelessed) version of the Dreambooth extension produce the best results.

I have also discovered that it is better to disable xformers for training and rather set memory attention to default, and that Torch version 1.13.1 works better than Torch version 2.

You can downgrade Torch as follows

cd stable-diffusion-webui
source venv/bin/activate
pip3 install torch==1.13.1 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu117
pip uninstall xformers
pip install https://huggingface.co/MonsterMMORPG/SECourses/resolve/main/xformers-0.0.19-cp310-cp310-manylinux2014_x86_64.whl

You should ensure that the versions of the Python dependencies in the requirements.txt file for the Dreambooth extension match the versions in the requirments.txt and requirements_versions.txt files of the Automatic1111 Web UI, and update the ones that don't match in the Automatic1111 Web UI, otherwise the Web UI will revert the dependencies each time its started and will result in a slow start-up time.

It is recommended to use a GPU instance with at least 20GB of VRAM, but preferably more to prevent GPU OOM errors, and you will need at least 32GB of memory for Dreambooth training, otherwise the Web UI will be OOM killed by the system kernel.

Step 1 - Clone the repository

git clone https://github.com/ashleykleynhans/dreambooth-api.git

Step 2 - Configure your Python Virtual Environment

Obviously this assumes that you already have Python installed on your system.

Change directory to the repository that you cloned above:

cd dreambooth-api

Create a new Python virtual environment:

python3 -m venv venv

Activate the Python virtual environment:

source venv/bin/activate

Install the required Python dependencies:

pip3 install -r requirements.txt

Step 3 - Update the configuration file

Edit config.yml and update the appropriate configuration settings.

Step 4 - Create a new Dreambooth model to train

You need to start of by creating a new Dreambooth model to train.

You can do this by running the script:

python3 create_model.py

Step 5 - Configure your newly created Dreambooth model

Once you have created your new Dreambooth model for training, you need to configure it.

You can do this by running the script:

python3 set_model_config.py

Step 6 - Start training your Dreambooth model

Once you have created your new Dreambooth model and configured it, you can proceed to train it.

You can do this by running the script:

python3 start_training.py

Training can be cancelled at any time, by running the cancel_training.py script (see Step 8 below).

Step 7 - Check Status

Once you have started training your new Dreambooth model, you can check the progress and get status images.

Step 7 (a) - Check Progress

You can check the progress of your training by running the script:

python3 get_status.py

Step 7 (b) - Get Status Images

You can do this by running the script:

python3 get_status_images.py

Step 8 - Cancel Training (Optional)

If you find that your sample images are over-trained, or if you want to cancel training for any other reason, you can do so by running the script:

python3 cancel_training.py