Modules
In Allor, modules are general components that are loaded into the ComfyUI, containing nodes that users interact with.
This page provides detailed information about each module, helping you understand their functionality.
In Allor, modules have the flexibility to utilize various implementations as a backend for nodes. This versatility allows for a diverse range of functionalities and interactions within the plugin.
Here is a list of backends:
PyTorch — Image processing with pure Tensor without transformations.
NumPy — PyTorch and NumPy arrays share underlying memory locations, allowing for efficient transformations without any performance loss.
OpenCV — An open-source library dedicated to real-time image processing. It offers even higher operation speeds than pure tensor processing.
Pillow — A widely used Python library for image manipulation. Slow speed of work, if possible, we will switch from it to OpenCV.
Rembg — A specialized library for removing image backgrounds. Despite its slower operation speed (up to 30 seconds), it provides effective background removal capabilities.
Alpha Chanel
A module for working with the transparency channel in images, enabling the creation, addition, restoration, or removal of the alpha channel.
Backend: PyTorch.
Image Container
This module allows the creation of an empty container based on images for use as a size value and for mathematical operations.
Backend: PyTorch.
Image Composite
This module enables the placement of images of varying sizes atop each other, preserving transparency.
Backend: PyTorch.
Image Segmentation
A module for removing backgrounds from images. For optimal results, it is recommended to add a prompt that generates a plain background.
Backend: Rembg.
Image Text
This module enables the creation of a text as an image. Plans include moving it to the Image Draw module.
Backend: Pillow.
Image Draw
A module for drawing figures as images. This module may undergo rework in the future.
Backend: Pillow.
Image Effects
A module for working with visual effects and optical distortions.
Image Filter
A module for altering images by applying mathematical operations. Unlike effects, it does not focus on artwork changes.
Backend: Pillow.
Image Noise
A module for working with noise on images. Currently, in the early stage of development.
Backend: NumPy.
Image Transform
A module for changing the representation of images. It enables resizing, reflecting, expanding, rotating, and cropping images.
Image Batch
A module for comfortable work with batch image processing. It allows adding images to a batch, splitting, removing, or getting the desired image from the array.
Backend: PyTorch.
Clamp
The module provides a strongly typed solution to improve the readability of nodes.
The nodes within this module uniquely have their module name positioned at the end of their name, rather than at the beginning.
Backend: PyTorch.