Forward-Forward and forward_forward_algorithm

These are **competitors**: both are independent reimplementations of the same algorithm (Hinton's Forward-Forward) in different frameworks (PyTorch vs TensorFlow), offering users a choice of backend rather than complementary functionality.

Forward-Forward
45
Emerging
Maintenance 0/25
Adoption 10/25
Maturity 16/25
Community 19/25
Maintenance 0/25
Adoption 7/25
Maturity 8/25
Community 15/25
Stars: 163
Forks: 28
Downloads:
Commits (30d): 0
Language: Python
License: MIT
Stars: 35
Forks: 6
Downloads:
Commits (30d): 0
Language: Jupyter Notebook
License:
Stale 6m No Package No Dependents
No License Stale 6m No Package No Dependents

About Forward-Forward

loeweX/Forward-Forward

Reimplementation of Geoffrey Hinton's Forward-Forward Algorithm

This project offers an alternative way to train deep neural networks that is inspired by how the brain learns. Instead of traditional methods, it processes 'real' data to enhance internal signals and 'fake' data to suppress them. This is useful for machine learning researchers and practitioners exploring novel deep learning architectures and training paradigms.

deep-learning-research neural-network-training biologically-inspired-AI machine-learning-algorithms

About forward_forward_algorithm

rmwkwok/forward_forward_algorithm

Implementation of Geoffrey Hinton's forward forward algorithm in tensorflow

This tool helps machine learning researchers explore and implement Geoffrey Hinton's Forward-Forward (FF) algorithm. It provides modular building blocks, like `FFLayer`s and a `TrainMgr`, to design and test different FF training sequences. Researchers working on novel neural network architectures or alternative training methods would use this to prototype and experiment.

neural-network-research deep-learning-algorithms unsupervised-learning alternative-training-methods machine-learning-experimentation

Scores updated daily from GitHub, PyPI, and npm data. How scores work