tinyCLAP: Distilling Constrastive Language-Audio Pretrained Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Paissan, Francesco, Farella, Elisabetta
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916407759863808
author Paissan, Francesco
Farella, Elisabetta
author_facet Paissan, Francesco
Farella, Elisabetta
contents Contrastive Language-Audio Pretraining (CLAP) became of crucial importance in the field of audio and speech processing. Its employment ranges from sound event detection to text-to-audio generation. However, one of the main limitations is the considerable amount of data required in the training process and the overall computational complexity during inference. This paper investigates how we can reduce the complexity of contrastive language-audio pre-trained models, yielding an efficient model that we call tinyCLAP. We derive an unimodal distillation loss from first principles and explore how the dimensionality of the shared, multimodal latent space can be reduced via pruning. TinyCLAP uses only 6% of the original Microsoft CLAP parameters with a minimal reduction (less than 5%) in zero-shot classification performance across the three sound event detection datasets on which it was tested
format Preprint
id arxiv_https___arxiv_org_abs_2311_14517
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle tinyCLAP: Distilling Constrastive Language-Audio Pretrained Models
Paissan, Francesco
Farella, Elisabetta
Sound
Computation and Language
Machine Learning
Audio and Speech Processing
Contrastive Language-Audio Pretraining (CLAP) became of crucial importance in the field of audio and speech processing. Its employment ranges from sound event detection to text-to-audio generation. However, one of the main limitations is the considerable amount of data required in the training process and the overall computational complexity during inference. This paper investigates how we can reduce the complexity of contrastive language-audio pre-trained models, yielding an efficient model that we call tinyCLAP. We derive an unimodal distillation loss from first principles and explore how the dimensionality of the shared, multimodal latent space can be reduced via pruning. TinyCLAP uses only 6% of the original Microsoft CLAP parameters with a minimal reduction (less than 5%) in zero-shot classification performance across the three sound event detection datasets on which it was tested
title tinyCLAP: Distilling Constrastive Language-Audio Pretrained Models
topic Sound
Computation and Language
Machine Learning
Audio and Speech Processing
url https://arxiv.org/abs/2311.14517