MMLForge Search models, datasets, tasks… Start free

Home › Models › Feature Extraction › MIT-SLS/USAD2-Base

U

MIT-SLS/USAD2-Base

Feature Extraction·MIT-SLS· 31

transformers cc-by-nc-sa-4.0 Feature Extraction 97.2M params

USAD 2.0: Scaling Representation Distillation for Universal Audio Understanding

Open in MLForge Sign up free Desktop app Source ↗

# pull & run locally
pip install mlforge-sdk && mlforge pull MIT-SLS/USAD2-Base

Model details

Task

Feature Extraction

Provider

MIT-SLS

Framework

transformers

Parameters

97.2M

Size

371 MB

License

cc-by-nc-sa-4.0

Downloads

31

Paper

arXiv:2203.03022

Updated

2026-06-05

About MIT-SLS/USAD2-Base

USAD 2.0: Scaling Representation Distillation for Universal Audio Understanding

Related Feature Extraction

B BAAI/bge-small-en-v1.5 Feature Extraction ·33.4M params 61.8M 497 🤗 HF B BAAI/bge-large-en-v1.5 Feature Extraction ·335.1M params 14.9M 688 🤗 HF Q Qwen/Qwen3-Embedding-0.6B Feature Extraction ·595.8M params 10.5M 1.1K 🤗 HF B BAAI/bge-base-en-v1.5 Feature Extraction ·109.5M params 8.6M 441 🤗 HF M intfloat/multilingual-e5-large Feature Extraction ·559.9M params 8.1M 1.2K 🤗 HF M mixedbread-ai/mxbai-embed-large-v1 Feature Extraction ·335.1M params 6.0M 810 🤗 HF

Browse all Feature Extraction models →