Fork, by Memoize Labs (V0.0.1.2)

Report bugs and get support at contact@memoizelabs.com

Installation

For first-time users, begin by running:

foo@user:~$ pip install memoizelabs

For upgrading to the latest version, upgrade to the latest version by running:

foo@user:~$ pip install --upgrade memoizelabs

Usage

Methods

Fork.fork(id, file_paths, isolated_voice_path, description='')

• Parameters:
  • id (str): the user-specified ID of the model.
  • file_paths (str array): a string of file paths containing objects to fine tune on. These can be .txt, .mp3, or .mp4 files.
  • isolated_voice_path (str): a path to a .mp3 file containing a clean recording of the target at least 30 seconds in length, used to pre-process submitted files and identify the target's voice within them before finetuning for better results.
  • description (str, optional): a description of the target (ex: An elderly woman with a deep, soothing voice, a British accent, and a witty sense of humor).

StateMachine.state()

• Returns:
  • The current state of the state machine, one of INIT, PRE-OP, OPERATIONAL, or FAULT.

StateMachine.errors()

• Returns:
  • Array of errors

Classes

Fork(api_key)

• Parameters:
  • api_key (str): your Memoize Labs Fork API key. Get one at memoizelabs.com/fork

StateMachine() : A state machine for using real time interactive features with low latency and continous connection to the server.

• States:
• INIT: the initialization state of the state machine. The state machine performs initial checks to confirm access to an audio device and wi-fi, then transitions to PRE-OP if checks pass or FAULT if an issue is detected.
• PRE-OP: the state machine is ready to begin live streaming of data upon command.
• OPERATIONAL: the state machine is connected and exchanging live streams with the server.
• FAULT: the fault state of the state machine. Once the causal errors are resolved, the user must transition the state machine back to INIT state by calling StateMachine.transition_state(INIT) and initialization checks will be re-run.