metadata

license: cc-by-4.0
library_name: timm

Model card for vit_base_patch16_1024_128.audiomae_as2m_ft_as20k

This is a port of AudioMAE ViT-B/32 weights for usage with timm. The naming convention is adopted from other timm's ViT models.

See the original repo here: https://github.com/facebookresearch/AudioMAE

A Vision Transformer (ViT) for audio. Pretrained on AudioSet-2M with Self-Supervised Masked Autoencoder (MAE) method, and fine-tuned on AudioSet-20k.

Model Details

Model Type: Audio classification / feature backbone
Papers:
- Masked Autoencoders that Listen: https://arxiv.org/abs/2207.06405
Pretrain Dataset: AudioSet-2M
Original: https://github.com/facebookresearch/AudioMAE

Model Usage

Audio Classification

from urllib.request import urlopen
import timm

# TODO: change this to audio
img = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model('gaunernst/vit_base_patch16_1024_128.audiomae_as2m_ft_as20k', pretrained=True)
model = model.eval()

# TODO: torchaudio.compliance.kaldi.fbank

output = model(transforms(img).unsqueeze(0))  # unsqueeze single image into batch of 1

top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5)

Audio Embeddings

from urllib.request import urlopen
import timm

# TODO: change this to audio
img = Image.open(urlopen(
    'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
    'gaunernst/vit_base_patch16_1024_128.audiomae_as2m_ft_as20k',
    pretrained=True,
    num_classes=0,  # remove classifier nn.Linear
)
model = model.eval()

# TODO: torchaudio.compliance.kaldi.fbank

output = model(transforms(img).unsqueeze(0))  # output is (batch_size, num_features) shaped tensor

# or equivalently (without needing to set num_classes=0)

output = model.forward_features(transforms(img).unsqueeze(0))
# output is unpooled, a (1, 197, 768) shaped tensor

output = model.forward_head(output, pre_logits=True)
# output is a (1, num_features) shaped tensor

Citation

@inproceedings{huang2022amae,
  title = {Masked Autoencoders that Listen},
  author = {Huang, Po-Yao and Xu, Hu and Li, Juncheng and Baevski, Alexei and Auli, Michael and Galuba, Wojciech and Metze, Florian and Feichtenhofer, Christoph}
  booktitle = {NeurIPS},
  year = {2022}
}

@misc{rw2019timm,
  author = {Ross Wightman},
  title = {PyTorch Image Models},
  year = {2019},
  publisher = {GitHub},
  journal = {GitHub repository},
  doi = {10.5281/zenodo.4414861},
  howpublished = {\url{https://github.com/huggingface/pytorch-image-models}}
}