Slimmable NAM: Neural Amp Models with adjustable runtime computational cost
This work addresses computational efficiency for musicians using neural amp models, but it is incremental as it adapts existing slimmable techniques to a specific domain.
The paper tackles the problem of high computational cost in neural amp models for musicians by introducing slimmable models that adjust size and cost without retraining, achieving negligible overhead and enabling real-time use in audio plugins.
This work demonstrates "slimmable Neural Amp Models", whose size and computational cost can be changed without additional training and with negligible computational overhead, enabling musicians to easily trade off between the accuracy and compute of the models they are using. The method's performance is quantified against commonly-used baselines, and a real-time demonstration of the model in an audio effect plug-in is developed.