MetaTF is used for the creation, training,
and testing of neural networks. MetaTF
supports the development of systems for
Edge AI on Brainchip’s Akida event
domain neural processor.

MetaTF Auto Converts
TensorFlow Models

The framework leverages the Python scripting language and ts associated tools and libraries, including Jupyter notebooks and NumPy.

  • Create models in TensorFlow Keras.
  • Convert using MetaTF easily and automatically.
  • No need to learn a new ML framework.
  • Convert your CNN to SNN with ease.

The MetaTF ML framework comprises three main python packages:

The Akida python package is an interface to the BrainChip Akida Neuromorphic System-on-Chip (NSoC). To allow the development of Akida models without actual Akida hardware, it includes a runtime, a Hardware Abstraction Layer (HAL), and a software backend that simulates the Akida NSoC.

The CNN2SNN tool provides means to convert Convolutional Neural Networks (CNN) that were trained using Deep Learning methods to event domain, low-latency and low-power network for use with the Akida runtime.

The Akida model zoo contains pre-created network models built with the Akida sequential API and the CNN2SNN tool using quantized Keras models.

MetaTF Development

MetaTF leverages the TensorFlow framework and PyPI for BrainChip tools installation.
The significant difference with other machine learning frameworks is that the data exchanged between layers is not the usual dense multidimensional arrays but sets of spatially organized events modeled as sparse multidimensional arrays.

Consultative Enablement

Use Cases

Data Set

with MetaTF

Optimize Network & Hardware Configuration