2022-10-30 15:15:09 +00:00
|
|
|
# HOWTO
|
|
|
|
|
|
2022-10-30 17:51:28 +00:00
|
|
|
This repo provide all the required config to run the Dreambooth version found in this note: https://note.com/kohya_ss/n/nee3ed1649fb6
|
2022-10-30 15:15:09 +00:00
|
|
|
|
2022-10-30 15:40:55 +00:00
|
|
|
|
2022-10-30 17:37:42 +00:00
|
|
|
## Required Dependencies
|
|
|
|
|
|
|
|
|
|
Python 3.10.6 and Git:
|
2022-10-30 17:39:00 +00:00
|
|
|
|
|
|
|
|
- Python 3.10.6: https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe
|
|
|
|
|
- git: https://git-scm.com/download/win
|
2022-10-30 17:37:42 +00:00
|
|
|
|
|
|
|
|
## Installation
|
|
|
|
|
|
2022-10-30 18:00:04 +00:00
|
|
|
Open a Powershell terminal and type the following inside:
|
|
|
|
|
|
|
|
|
|
```powershell
|
2022-10-30 15:19:24 +00:00
|
|
|
git clone https://github.com/bmaltais/kohya_ss.git
|
|
|
|
|
cd kohya_ss
|
2022-10-30 15:15:09 +00:00
|
|
|
python -m venv --system-site-packages venv
|
|
|
|
|
.\venv\Scripts\activate
|
|
|
|
|
|
|
|
|
|
pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 --extra-index-url https://download.pytorch.org/whl/cu116
|
|
|
|
|
|
|
|
|
|
pip install --upgrade diffusers
|
|
|
|
|
pip install -r requirements.txt
|
|
|
|
|
pip install OmegaConf
|
|
|
|
|
pip install pytorch_lightning
|
|
|
|
|
|
|
|
|
|
pip install -U -I --no-deps https://github.com/C43H66N12O12S2/stable-diffusion-webui/releases/download/f/xformers-0.0.14.dev0-cp310-cp310-win_amd64.whl
|
|
|
|
|
|
2022-10-30 15:40:55 +00:00
|
|
|
# Setup bitsandbytes with Adam8bit support for windows: https://note.com/kohya_ss/n/n47f654dc161e
|
2022-10-30 15:15:09 +00:00
|
|
|
pip install bitsandbytes==0.35.0
|
|
|
|
|
cp .\bitsandbytes_windows\*.dll .\venv\Lib\site-packages\bitsandbytes\
|
|
|
|
|
cp .\bitsandbytes_windows\cextension.py .\venv\Lib\site-packages\bitsandbytes\cextension.py
|
|
|
|
|
cp .\bitsandbytes_windows\main.py .\venv\Lib\site-packages\bitsandbytes\cuda_setup\main.py
|
|
|
|
|
|
2022-10-30 15:18:20 +00:00
|
|
|
accelerate config:
|
|
|
|
|
- 0
|
|
|
|
|
- 0
|
2022-10-30 15:40:55 +00:00
|
|
|
- NO
|
|
|
|
|
- NO
|
2022-10-30 15:18:20 +00:00
|
|
|
- All
|
2022-10-30 15:40:55 +00:00
|
|
|
- fp16
|
2022-10-30 17:37:42 +00:00
|
|
|
```
|
|
|
|
|
|
2022-10-30 17:51:28 +00:00
|
|
|
## Folders configuration
|
|
|
|
|
|
|
|
|
|
Refer to the note to understand how to create the folde structure. In short it should look like:
|
|
|
|
|
|
|
|
|
|
```
|
|
|
|
|
<wathever top folder name>
|
|
|
|
|
|- reg_<class>
|
|
|
|
|
|- <repeat count>_<prompt>
|
|
|
|
|
|- train_<class>
|
|
|
|
|
|- <repeat count>_<prompt>
|
|
|
|
|
```
|
|
|
|
|
|
|
|
|
|
Example for `sks dog`
|
|
|
|
|
|
|
|
|
|
```
|
|
|
|
|
my_sks_dog_dreambooth
|
|
|
|
|
|- reg_dog
|
|
|
|
|
|- 1_sks dog
|
|
|
|
|
|- train_dog
|
|
|
|
|
|- 20_sks dog
|
|
|
|
|
```
|
|
|
|
|
|
2022-10-30 17:37:42 +00:00
|
|
|
## Execution
|
2022-10-30 15:15:09 +00:00
|
|
|
|
2022-10-30 18:01:12 +00:00
|
|
|
Edit and paste the following in a Powershell terminal:
|
|
|
|
|
|
2022-10-30 18:00:04 +00:00
|
|
|
```powershell
|
2022-10-30 15:15:09 +00:00
|
|
|
accelerate launch --num_cpu_threads_per_process 6 train_db_fixed_v6.py `
|
2022-10-30 18:06:20 +00:00
|
|
|
--pretrained_model_name_or_path="D:\models\last.ckpt" `
|
2022-10-30 17:37:42 +00:00
|
|
|
--train_data_dir="D:\dreambooth\train_bernard\train_man" `
|
2022-10-30 18:06:20 +00:00
|
|
|
--reg_data_dir="D:\dreambooth\train_bernard\reg_man" `
|
2022-10-30 17:37:42 +00:00
|
|
|
--output_dir="D:\dreambooth\train_bernard" `
|
2022-10-30 15:15:09 +00:00
|
|
|
--prior_loss_weight=1.0 `
|
|
|
|
|
--resolution=512 `
|
|
|
|
|
--train_batch_size=1 `
|
|
|
|
|
--learning_rate=1e-6 `
|
2022-10-30 17:37:42 +00:00
|
|
|
--max_train_steps=2100 `
|
2022-10-30 15:15:09 +00:00
|
|
|
--use_8bit_adam `
|
|
|
|
|
--xformers `
|
|
|
|
|
--mixed_precision="fp16" `
|
|
|
|
|
--cache_latents `
|
|
|
|
|
--gradient_checkpointing `
|
2022-10-30 17:37:42 +00:00
|
|
|
--save_every_n_epochs=1
|
