Prerequisites

Before installing DiffEngine, please ensure that PyTorch has been successfully installed following the official guide.

Installation

Below are quick steps for installation:

pip install git+https://github.com/okotaku/diffengine.git

Get Started

DiffEngine makes training easy through its pre-defined configs. These configs provide a streamlined way to start your training process. Here’s how you can get started using one of the pre-defined configs:

  1. Choose a config: You can find various pre-defined configs in the configs directory of the DiffEngine repository. For example, if you’re interested in training a model for DreamBooth using the Stable Diffusion algorithm, you can use the configs/stable_diffusion_dreambooth/stable_diffusion_v15_dreambooth_lora_dog.py.

  2. Start Training: Open a terminal and use the following command to start training with the chosen config:

diffengine train stable_diffusion_v15_dreambooth_lora_dog

  1. Monitor Progress and get results: The training process will begin, and you will see output similar to the provided example. You can monitor the progress of your training as it proceeds. The outputs of training is in work_dirs/${CONFIG_NAME}, work_dirs/stable_diffusion_v15_dreambooth_lora_dog in this case.

work_dirs/stable_diffusion_v15_dreambooth_lora_dog
├── 20230802_033741
|   ├── 20230802_033741.log  # log file
|   └── vis_data
|         ├── 20230802_033741.json  # log json file
|         ├── config.py  # config file for each experiment
|         └── vis_image  # visualized image from each step
├── step1199
|   └── pytorch_lora_weights.bin  # weight for inferencing with diffusers.pipeline
├── iter_1200.pth  # checkpoint from each step
├── last_checkpoint  # last checkpoint, it can be used for resuming
└── stable_diffusion_v15_dreambooth_lora_dog.py  # latest config file

An illustrative output example is provided below:

img

Example Notebook

Open In Colab

For a more hands-on introduction to DiffEngine, you can run the Example Notebook on Colaboratory. This notebook demonstrates the process of training using SDV1.5 and SDV2.1 DreamBooth configurations.

Docker

Below are quick steps for installation and run dreambooth training by docker:

git clone https://github.com/okotaku/diffengine
cd diffengine
docker compose up -d
docker compose exec diffengine train stable_diffusion_v15_dreambooth_lora_dog