charloratools

0

Easier management of image directories with python.

Productivity

python3
python-package
lora-training
computer-vision

Charloratools Logo

version build tests codecov

This is a simple python library which intends to facilitate the process of image gathering,filtering and processing when creating a training dataset to use for training stable diffusion LoRas. As the name suggests it is specifically useful when organizing a character lora image training dataset.

Installation

<a href=https://pypi.org/project/charloratools/><img src=https://imagedelivery.net/YCQ3OFRYiR1R_AeUslNHiw/9e17a46e-346c-4475-11f3-fd0d661c1800/400x400><a href=https://pypi.org/project/charloratools/>

pip install charloratools

Torch CPU-Only

pip install charloratools[full]

Installation Steps

Torch is not installed during installation through pip due to specific installation requirements for different OSes. This package won't work unless you manually install those.

You can also choose to install the OS independent version of torch ( uses only the CPU ) by running the "Torch CPU-Only" pip command listed above.

Dependency Install Script

After installing with pip you can run a script that automatically detects which distribution of torch is the correct one for your current machine and installs it.

The script can also be used to check if you have the correct distribution of torch installed or to install the latest correct distribution for your OS.

The script checks if you already have torch installed and quits without making any changes if you do.

Installing and Running The Script

pip install charloratools
charloratools install_torch

Installing Torch Manually

If you don't want to run the script and prefer installing torch yourself I recommend following the steps described in their website. Once torch is installed the package should work correctly.

ComfyUI Custom Node

Some of the functionality of this package is already available as a custom node for ComfyUI (comfy-ui-lora-dataset-tools) , I'll be implementing the rest of the functionalities in new versions of the custom node.

Roadmap

  • Create better documentation in wiki