Neural Network

The Neural Network enables the deployment of real-time deep learning algorithms, in a .linn format, into your Multi-Instrument setup. See the examples and documentation for how to write and export your neural network .linn file in Python.

If you are directly using the RESTful API (e.g. using cURL), the instrument name as used in the URL is neuralnetwork.

Multi-instrument Mode

Moku Neural Network instruments can only be used in Multi-instrument Mode. Refer to the Multi-instrument Mode Getting Started Guide for more details.

Loading your .linn file

Before you can deploy a Moku Neural Network instrument, you must load the .linn file that you have generated in Python onto your Moku. When using the API, this can be accomplished through the upload_network function.

Functions

Core

• Multi-instrument: Neural Network in multi-instrument context
• describe_network: Provide a description of the currently loaded network
• load_settings: Load a previously saved `.mokuconf` settings file into the instrument. To create a `.mokuconf` file, either use `save_settings` or the desktop app.
• save_settings: Save instrument settings to a file. The file name should have a `.mokuconf` extension to be compatible with other tools.
• set_defaults: Set the Neural Network to its default state
• set_input: Set the voltage range for a given input
• set_input_sample_rate: Set the input sample rate
• set_output: Set the output range for a given output
• summary: Returns a short summary of current instrument state
• upload_network: Upload a neural network in .linn JSON format