Convert loss-friendly output into actual symbolic transcription. Parameters ---------- raw_output : Tensor (B x T x

Convert loss-friendly output into actual symbolic transcription. Parameters ---------- raw_output : Tensor (B x T x

Determine the minimum and maximum pitch for each slice of stacked notes. Parameters ---------- stacked_notes : dict Dictionary

Perform the main processing steps for TabCNN. Parameters ---------- feats : Tensor (B x T x C x F x W) Inp

Perform the main processing steps for the transcription model. Parameters ---------- feats : Tensor (B x ...)

Perform the main processing steps for the output layer. Parameters ---------- feats : Tensor (see child class for ex

Perform the main processing steps for the softmax groups. Parameters ---------- feats : Tensor (B x T x F)

Perform the main processing steps for the output layer. Parameters ---------- feats : Tensor (B x T x F) I

Perform the main processing steps for Onsets & Frames (V2). Parameters ---------- feats : Tensor (B x C x T x F)

Feed features through the acoustic model. Parameters ---------- in_feats : Tensor (B x C x T x F) Input fe

Feed features through the music language model. Parameters ---------- in_feats : Tensor (B x T x E) Input

Feed features through the music language model. Parameters ---------- in_feats : Tensor (B x 1 x E) Input

Chunk activations into overlapping frames along the last dimension. Parameters ---------- activations : ndarray Provided activ

Obtain the estimator's active notes as a stacked multi pitch array. Returns ---------- stacked_multi_pitch : ndarray

Obtain the estimator's active notes as stacked notes. Parameters ---------- current_time : float or None (Optional)

Obtain information about the device currently chosen. Returns ---------- device : dict Device information

Default key for multi pitch activations.

Default key for note estimates.

Default key for pitch lists.

Default key for onsets activation maps.

Default key for offsets activation maps.

This should not be called directly on a ComboEvaluator.

Default key for loss.

Default key for multi pitch activation maps.

Default key for notes.

Default key for pitch lists.

Default key for tablature.

Default key for tablature.

Obtain the instrument's range of MIDI pitches across each degree of freedom. Returns ---------- pitch_ranges : ndarr

Determine the number of frames the module will return for a piece of audio. Parameters ---------- audio : nd

Determine the number of frames the module will return for a piece of audio. Parameters ---------- audio : nd

Helper function to access dimensionality of features. Returns ---------- feature_size : int Dimensionality

Helper function to access dimensionality of features. Returns ---------- feature_size : int Dimensionality

Helper function to access dimensionality of features. Returns ---------- feature_size : int Dimensionality

Helper function to access dimensionality of features. Returns ---------- feature_size : int Dimensionality

Helper function to access dimensionality of features.

Determine the start time of each frame for given audio parameters Parameters ---------- duration : float Total length (secon

Helper function to access hop length. Returns ---------- hop_length : int or float Number of samples betwe

Compute the cross entropy softmax loss for each group independently. Parameters ---------- estimated : Tensor (B x T

Compute the binary cross entropy loss for each key independently. Parameters ---------- estimated : Tensor (B x T x

Get the path to the annotations of a track. Parameters ---------- track : string MAESTRO track name

Helper function to access number of feature channels. Returns ---------- num_channels : int Number of inde

Determine how many frets are supported by this profile. Returns ---------- num_frets : int Number of frets

Get the MIDI pitch for a given string and fret. Parameters ---------- string : int String to query

Obtain indices to resample from a set of original times to a set of target times using nearest neighbor interpolation. Parameters --

Determine the range of audio samples which will produce features with a given number of frames. This will inevitably fall ont

Determine the range of audio samples which will produce features with a given number of frames. Parameters ---------

Determine the range of audio samples which will produce features with a given number of frames. Parameters ---------

Helper function to access sampling rate. Returns ---------- sample_rate : int or float Presumed sampling r

Simple helper function to create a tag for saving a file if one does not already exist. This is useful because in some places we don't know

Determine the time, in seconds, associated with each frame. Parameters ---------- audio : ndarray Mono-cha

Determine the time, in seconds, associated with each frame. Parameters ---------- audio: ndarray Mono-chan

Determine the time, in seconds, associated with each frame. This is the default behavior. It can be overridden. Parameters

Determine the time, in seconds, associated with frame. Parameters ---------- audio: ndarray Mono-channel a

Parent class MAPS has and uses this method. It is overridden here to squash the MAPS functionality if the function were called from

Get the tracks associated with a dataset partition. Parameters ---------- split : string Name of the parti

Get the tracks associated with a dataset partition. Parameters ---------- split : string Name of the parti

Get all the possible fretboard positions for a given MIDI pitch. Parameters ---------- midi_pitch : int MI

Get the path to the audio of a track. Parameters ---------- track : string MAESTRO track name Ret

Change default matplotlib parameter so no toolbar is added to plots.

Remove any activations within a specified time window following a previous activation. TODO - this is extremely slow for non-sparse activati

Linearly interpolate between gaps in a one-dimensional array. Parameters ---------- arr : ndarray One-dimensional array ga

Load the ground-truth from memory or generate it from scratch. Parameters ---------- track : string Name o

Simple helper function to load a NumPy zip file. Parameters ---------- path : string Path to load the NumPy zip file Retu

Helper function to load a JAMS file and extract the duration. Parameters ---------- jams_path : string Path to JAMS file to re

Load audio from a file and normalize it. Parameters ---------- wav_path : string Path to audio file to read fs : int or fl

Helper function to load a JAMS file and extract the notes. Parameters ---------- jams_path : string Path to JAMS file to read

Load all MIDI notes from a MIDI file, keeping track of sustain pedal activity. TODO - make sustain pedal stuff optional? TODO - break thi

Helper function to load a JAMS file and extract a pitch list. Parameters ---------- jams_path : string Path to JAMS file to re

Helper function to load a JAMS file and extract the stacked notes. Parameters ---------- jams_path : string Path to JAMS file

Helper function to load a JAMS file and extract a stacked pitch list. Parameters ---------- jams_path : string Path to JAMS fi

View logistic activations as a stacked multi pitch array. Parameters ---------- logistic : ndarray or tensor (..., N x T) Arra

View logistic activations as tablature class membership. Parameters ---------- logistic : tensor (..., N x T) Array of distinc

Decode a multi pitch array into loose MIDI note groups. Parameters ---------- multi_pitch : ndarray (F x T) Discrete pitch act

Convert a collection of notes into a dictionary of stacked notes. Parameters ---------- pitches : ndarray (N) Array of pitches

Add a semitone offset to note groups. Parameters ---------- pitches : ndarray (N) Array of pitches corresponding to notes

Function called when a key is pressed. Parameters ---------- key : pynput.keyboard.Key Key which was press

Function called when a key is released. Parameters ---------- key : pynput.keyboard.Key Key which was rele

Package together the loose times and observations of a pitch list in a save-friendly format. Parameters ---------- times : ndarray (

Package together the key-value pairs of a stacked representation in a save-friendly format. Parameters ---------- stacked_representa

Static function for plotting note groups. TODO - make compatible w/ real-time note plotting like other visualization functions Paramete

Calculate loss and finalize model output. Parameters ---------- batch : dict Dictionary including model ou

Calculate loss and finalize model output. Parameters ---------- batch : dict Dictionary including model ou

Perform necessary pre-processing steps for the transcription model. Parameters ---------- batch : dict Dic

Perform necessary pre-processing steps for the transcription model. Parameters ---------- batch : dict Dic

Add a specified amount of empty frames to the buffer. Parameters ---------- amount : int Number of empty f

Get the VQT features stacked across harmonics for a piece of audio. Parameters ---------- audio : ndarray

Chop the audio in frames according to window and hop length. Parameters ---------- audio : ndarray Mono-ch

Get the signal power for each frame of a piece of audio. Parameters ---------- audio : ndarray Mono-channe

Get the spectrogram features for a piece of audio. Parameters ---------- audio : ndarray Mono-channel audi

Get the Mel Spectrogram features for a piece of audio. Parameters ---------- audio : ndarray Mono-channel

Get features for a piece of audio. Parameters ---------- audio : ndarray Mono-channel audio

Get the features for a piece of audio. Parameters ---------- audio : ndarray Mono-channel audio R

Process the track independently using each estimator. Parameters ---------- raw_output : dict Dictionary c

Very similar to parent method, except file is written after results are calculated for each evaluator and packaged into a single dict

Determine if the stream has finished.

Remove blips (single-frame positives) in activations. Parameters ---------- activations : ndarray Provided activations Re