cral 0.4.0

CRAL: Library for CNNs

CNN Research Abstraction Library

The CNN Research Abstraction Library or CRAL in short is a deep learning computer vision library for data scientists, researchers, and developers. With a primary focus on applied deep learning, the CRAL library encourages rapid development and comes with ready-to-use state-of-the-art networks and other pragmatic tools for a variety of applications in the computer vision space.

Our aim is also to make it easier to reproduce and extend the results of various Deep Learning-powered Computer Vision (DLCV) algorithms developed in academia and industrial labs.

List of Algorithms

Object detection

• RetinaNet
• yolov3
• SSD
• FasterRCNN

Instance Segmentation

• MaskRCNN

Semantic Segmentation

• UNet
• UNet ++
• Deeplabv3+
• FpnNet
• PspNet
• SegNet
• LinkNet

Guiding Principles

Simple: To make it easy for deep learning engineers & students alike to use neural networks to build computer vision applications of their choice, using low code approach.

Fast: To accelerate going from experimentation to a working model.

Reproducible: To offer implementations that can easily be trained and reproduced on your own data.

Components

CRAL has a modular design to enable you to use each of its components independently, Alternatively, you can use the pipeline to get started quickly with multiple networks out-of-the-box.

Components	Description
CNN models	Ready to use implementations of State-of-the-art (SOTA) algorithms.
Pipeline tools	Load and validate your data before you start training.
Optimization and debugging	Integration with Experiment Tracking, HP Optimization and other toolsets to help faster and build transparent models

Detailed documentation: Link