https://github.com/techpro-studio/nntoolkitcore

Cross-platform mobile Neural network C library for training and inference on the device. CPU only. It fits for time-series data.
https://github.com/techpro-studio/nntoolkitcore

accelerate android audio-processing batchnorm conv1d dense dft gru ios lstm macos neural-network nn-filters rnn spectrogram watchos

Last synced: 10 months ago
JSON representation

Cross-platform mobile Neural network C library for training and inference on the device. CPU only. It fits for time-series data.

• Host: GitHub
• URL: https://github.com/techpro-studio/nntoolkitcore
• Owner: techpro-studio
• License: mit
• Created: 2020-09-30T12:50:25.000Z (over 5 years ago)
• Default Branch: master
• Last Pushed: 2021-10-17T08:19:09.000Z (over 4 years ago)
• Last Synced: 2025-03-24T03:17:11.571Z (11 months ago)
• Topics: accelerate, android, audio-processing, batchnorm, conv1d, dense, dft, gru, ios, lstm, macos, neural-network, nn-filters, rnn, spectrogram, watchos
• Language: C
• Homepage:
• Size: 212 KB
• Stars: 8
• Watchers: 0
• Forks: 1
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# NNToolkitCore

Cross-platform mobile Neural Network library for training and inference on device.

CPU only.

On Apple platforms it uses Accelerate framework as a backend;

On other platforms it uses Eigen for matmul, kissfft for dft, and ARM NEON intrinsics for acceleration.

NN layers:

Conv1d

GRU

RNN

LSTM

BatchNorm

Activation

TimeDistributedDense

Dense

DSP tools:

Spectrogram

Window functions

Algorithms were tested using tf 2.3.0

Examples and tests will be exposed in different repository.