Data Preparation

Convert audio files and annotations into training-ready format by splitting audio into clips, matching annotations, and creating positive/negative class labels.

$> # Prepare the project's training data
$> koogu-prepare my_project.config

The preprocessing step will:

  1. Pre-process audio files by applying necessary filtering and splitting into fixed-duration segments

  2. Match annotations to segments based on temporal overlap

  3. Label each segment as positive or negative for each class

  4. Save prepared segments and annotaions in a format optimized for training

Parameters

Positional arguments

  • <CONFIG FILE>

    Path to config file.

Input control

  • --filetypes EXTN [EXTN ...]

    Audio file types to restrict processing to. Can specify multiple types separated by whitespaces. By default, will process all discovered files with the following extensions: [.wav, .WAV, .flac, .aif, .mp3]. When train_audio_annotations_map is available, this option is only considered where audio sources are directories.

Process control

  • --threads NUM

    Number of threads to spawn for parallel execution.

    Default: as many CPUs

Logging

  • --log LOGFILE

    If specified, logging will be written out to this file instead of the default.

    Default: PROJECT-LOGS-DIR/prepare.log

  • --loglevel LEVEL

    Logging level. Choices: CRITICAL, ERROR, WARNING, INFO, DEBUG.

    Default: INFO