Neural Python Packages

neuralprophet

NeuralProphet: A simple forecasting package

neuralpiano

[WIP] Hi-Fi neural MIDI piano synthesizer and MIDI renderer

neuralnet-visualize

Give architecture get the image

torchsample

Neural network sampling utilities.

atiny

Atiny is a gradient based PyTorch optimizer. Atiny uses half of Adam’s memory to achieve results that are not inferior to Adam.

tempocnn

Framework for estimating temporal properties of music tracks.

neural-tools

Luminance and histogram matching tools

okrolearn

A machine learning library created in python by Okerew

fwu

FWU is a series of PyTorch-based neural network units. It has higher information utilization, prevents neuronal necrosis, and eliminates the need to add additionall activation functions.

glassinit

glassInit is a way to initialize the parameters of a neural network layer @pytorch. It is suitable for shortcut connections where tensor sizes need to be adjusted. The flow of information between forward and backward propagation is maximized while adjusting the tensor size.

orientloss

orientLoss is a classification loss function @PyTorch. It less prone to overfitting compared to crossEntropy. Compatible with random vector/embeddings with direction-indicating categories、probabilistic、oneHot as target.

titans-memory

PyTorch implementation of Titans: Learning to Memorize at Test Time (Behrouz, Zhong & Mirrokni, 2024)

haloblocks

Modern, modular, composability-first neural network building blocks for PyTorch.

neurokernel

Neurokernel Project

inr-collection

A collection of conditional \ modulatable implicit neural representation implementations and basic building blocks in PyTorch.

promptscan

AI-powered prompt injection detective - detect and analyze suspicious prompts

deepevolution

A genetic algorithm to evolve tensorflow keras neural networks.

jax-dips

JAX-DIPS is a differentiable interfacial PDE solver.

fftloss

The fftLoss @PyTorch is a frequency domain loss function that prevents the problem of weak frequency components being suppressed by strong frequency components when using a regular loss function.

ttscube

End-2-end speech synthesis with recurrent neural networks