ADTOF Frame_RNN (ONNX)

Automatic drum transcription of full music mixes: bass drum, snare, toms, hi-hat, cymbals and ride. A conversion to ONNX of the Frame_RNN model of ADTOF (Zehren, Alunno, Bientinesi), trained on 359 hours of crowdsourced, non-synthetic music. Non-commercial use only (CC BY-NC-SA 4.0).

Input waveform: [1, samples] float32, mono, 44.1 kHz, values in [-1, 1]. The log-filtered spectrogram (madmom's, 84 bands, 100 frames per second) is part of the graph.
Output activations: [1, ceil(samples / 441), 5], sigmoid activations at 100 frames per second for bass drum, snare, toms, hi-hat and cymbals with ride.
Hits come from peak picking per class with the authors' thresholds 0.22, 0.24, 0.32, 0.22, 0.30 (madmom NotePeakPickingProcessor, pre_avg=0.1, post_avg=0.01, pre_max=0.02, post_max=0.01, combine=0.02).
Matches the authors' Keras model fed with madmom features to within 2e-6.

Converted by the Silly MIDI Tools scripts in scripts/drums/.

See NOTICE.md for the attribution and the changes made, and LICENSE for the licence text.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support