An attention model improves multi-label instrument recognition accuracy on the weakly labeled OpenMIC dataset compared to baseline, RNN, and fully connected networks across 20 instruments.
Here, the input features of dimension 10× 128 are flattened into a single feature vector for classification
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.IR 1years
2019 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
An Attention Mechanism for Musical Instrument Recognition
An attention model improves multi-label instrument recognition accuracy on the weakly labeled OpenMIC dataset compared to baseline, RNN, and fully connected networks across 20 instruments.