Model Card for esp-aves2-sl-eat-all-ssl-all

Model Details

Model Description

esp-aves2-sl-eat-all-ssl-all is an audio representation learning model (bioacoustic encoder) trained with a two-stage recipe: self-supervised pretraining of EAT on the All mix (Bio + AudioSet), followed by supervised post-training on the same All mix, as described in What Matters for Bioacoustic Encoding.

• Developed by: Marius Miron, David Robinson, Milad Alizadeh, Ellen Gilsenan-McMahon, Gagan Narula, Emmanuel Chemla, Maddie Cusimano, Felix Effenberger, Masato Hagiwara, Benjamin Hoffman, Sara Keen, Diane Kim, Jane K. Lawton, Jen-Yu Liu, Aza Raskin, Olivier Pietquin, Matthieu Geist
• Funded by: More info at https://www.earthspecies.org/about-us#support
• Shared by: Earth Species Project
• Model type: Transformer; EAT backbone
• License: CC-BY-NC-SA
• Finetuned from model: EAT-all (SSL) (see Parent Models)

Model Sources

• Repository: https://github.com/earthspecies/avex
• Paper: What Matters for Bioacoustic Encoding
• Hugging Face Model: ESP-AVES2 Collection
• Configuration: train_config.yaml

Parent Models

1. EAT (Efficient Audio Transformer)
   • Source: http://github.com/cwx-worst-one/EAT
   • Description: Open-source transformer audio encoder; the paper uses it to study modifications to SSL pretraining and subsequent supervised post-training.
   • License: See upstream repository

Uses

Direct Use

esp-aves2-sl-eat-all-ssl-all can be used as an embedding model for bioacoustic tasks such as species classification/detection, retrieval and clustering, individual ID, and repertoire analysis.

Downstream Use

Use frozen embeddings with linear probes, or fine-tune on your target task/domain.

Out-of-Scope Use

Not a generative model; does not output text.

Bias, Risks, and Limitations

• Bias: Training data biases (taxa, geography, recording conditions) can affect downstream performance.
• Risks: Potential misuse for harmful wildlife exploitation; apply safeguards.
• Limitations: 16 kHz standardization in the paper; may not capture higher-frequency information important for some taxa.

How to Get Started with the Model

Loading this model requires the AVEX (Animal Vocalization Encoder) library avex to be installed.

Installation

pip install avex

Or with uv:

uv add avex

For more details, see https://github.com/earthspecies/avex.

Loading the Model

from avex import load_model

model = load_model("esp_aves2_sl_eat_all_ssl_all", device="cuda")

Using the Model

# Case 1: embedding extraction (features only)
backbone = load_model("esp_aves2_sl_eat_all_ssl_all", device="cuda", return_features_only=True)

with torch.no_grad():
    embeddings = backbone(audio_tensor)
    # Shape: (batch, time_steps, 768) for EAT

# Pool to get fixed-size embedding
embedding = embeddings.mean(dim=1)  # Shape: (batch, 768)

# Case 2: supervised predictions (logits over label IDs; see label_map.json)
model = load_model("esp_aves2_sl_eat_all_ssl_all", device="cuda")

with torch.no_grad():
    logits = model(audio_tensor)
    predicted_class = logits.argmax(dim=-1).item()

Transfer Learning with Probes

from avex.models.probes import build_probe_from_config
from avex.configs import ProbeConfig

# Load backbone for feature extraction
base = load_model("esp_aves2_sl_eat_all_ssl_all", return_features_only=True, device="cuda")

# Define a probe head for your task
probe_config = ProbeConfig(
    probe_type="linear",
    target_layers=["last_layer"],
    aggregation="mean",
    freeze_backbone=True,
    online_training=True,
)

probe = build_probe_from_config(
    probe_config=probe_config,
    base_model=base,
    num_classes=10,  # Your number of classes
    device="cuda",
)

Class Label Mapping

The class label mapping for this supervised learning model can be found at label_map.json in the Hugging Face repository.

Training Details

Training Data

esp-aves2-sl-eat-all-ssl-all uses the paper’s two-stage recipe:

• SSL pretraining: EAT on All (Bio + AudioSet) to produce EAT-all
• Supervised post-training: on All to produce sl-EAT-all

Training Data Sources

Dataset	Description	Source	License	Size
AudioSet	general audio	Link	See dataset terms	5700 hours
Xeno-canto	birds	Link	CC (varies)	10416 hours
iNaturalist	diverse taxa	Link	CC (varies)	1539 hours
Watkins	marine mammals	Link	licensing agreement (paper)	27 hours
Animal Sound Archive	diverse taxa	Link	See archive terms	78 hours

Training Procedure

As described in the paper:

• SSL objective: a mix of teacher distillation and reconstruction of masked spectrogram patches.
• Augmentations: random additive noise (p=0.5, SNR in ([-10, 20]) dB); mixup-style within-batch mixing (p=0.5) with union of labels during supervised post-training.

Training Hyperparameters

Training hyperparameters are specified in train_config.yaml.

Evaluation

Testing Data, Factors & Metrics

Testing Data

The paper evaluates on:

• BEANS (classification and detection): https://github.com/earthspecies/beans
• BirdSet (detection): https://huggingface.co/datasets/DBD-research-group/BirdSet
• Individual ID: Pipit, Chiffchaff, Little Owl, Macaques
• Vocal Repertoire: Zebra Finch, Giant Otters, Bengalese Finch, Killer Whale

Metrics

• Linear probing: accuracy / mAP
• Retrieval: ROC AUC
• Clustering: NMI

Results

Aggregate results for linear probing (frozen base model) with esp-aves2-sl-eat-all-ssl-all (from the provided LaTeX table):

Benchmark	Task	Metric	Score
BEANS Classification	Probe	Accuracy	0.788
BEANS Classification	Retrieval	ROC AUC	0.791
BEANS Classification	Clustering	NMI	0.536
BEANS Detection	Probe	mAP	0.356
BEANS Detection	Retrieval	ROC AUC	0.704
BirdSet	Probe	mAP	0.255
BirdSet	Retrieval	ROC AUC	0.706
Individual ID	Probe	Accuracy	0.456
Individual ID	Retrieval	ROC AUC	0.637
Vocal Repertoire	Retrieval	ROC AUC	0.798
Vocal Repertoire	Clustering	NMI	0.530

Citation

BibTeX:

@inproceedings{miron2025matters,
  title={What Matters for Bioacoustic Encoding},
  author={Miron, Marius and Robinson, David and Alizadeh, Milad and Gilsenan-McMahon, Ellen and Narula, Gagan and Chemla, Emmanuel and Cusimano, Maddie and Effenberger, Felix and Hagiwara, Masato and Hoffman, Benjamin and Keen, Sara and Kim, Diane and Lawton, Jane K. and Liu, Jen-Yu and Raskin, Aza and Pietquin, Olivier and Geist, Matthieu},
  booktitle={The Fourteenth International Conference on Learning Representations},
  year={2026}
}

Model Card Contact

Contact: marius@earthspecies.org, david@earthspecies.org, milad@earthspecies.org, gagan@earthspecies.org