stratus-labs/nocturne-v1-teacher

Nocturne v1.1 (Teacher) — Bioacoustic Species Recognition for Non-Bird Taxa

Nocturne is a bioacoustic species classifier trained by Stratus Labs covering the taxa that BirdNET and Perch don't: insects, amphibians, non-bird mammals, and reptiles. It's the "night side" of the soundscape — the taxa that are dominantly nocturnal or crepuscular, whose acoustic signal is the half of biodiversity monitoring bird-focused models leave behind.

v1.1 (this revision) uses a 3× higher learning rate than v1 during backbone fine-tuning. Independent runs at both LRs plateaued around macro-F1 0.10; v1.1 edged past by ~3 %. Prior v1 weights are preserved in the commit history if you want to pin to revision=<v1-commit-sha>.

Try it live (no install)

Live public endpoint hosted on Stratus Labs infra (Apple M4 Max, MPS-backed):

• UI: https://nocturne.runstratus.com/ — drag-and-drop upload
• API:

curl -X POST "https://nocturne.runstratus.com/predict?top_k=5&threshold=0.15" \
     -F "file=@your_clip.wav"

Rate-limited to 10 requests per minute per IP. Uploads capped at 25 MB / 5 minutes per request.

Model at a glance

Backbone	Audio Spectrogram Transformer (AST), initialized from MIT/ast-finetuned-audioset-10-10-0.4593
Head	Linear over 2182 species (multi-label BCE)
Input	10-second mono waveform @ 16 kHz -> 128-band log-mel
Params	~86M
Precision	bf16 for training and inference
Trained on	1x NVIDIA GB10 (DGX Spark), unified 128 GB memory
License (weights)	CC-BY-4.0
License (code)	Apache-2.0

See stratus-labs/nocturne-v1-mini for a distilled 7M-parameter EfficientNet-B1 student for CPU / edge deployment.

Evaluation

Real numbers, run on the held-out splits. Earlier drafts of this card cited macro-F1 at threshold 0.5 (too high — kills recall). These are the corrected numbers at threshold 0.3 (default) and with per-class thresholds calibrated on val:

split	metric	threshold 0.3 (default)	calibrated per-class
val (13,391 samples)	macro-F1	0.139	0.184
	micro-F1	0.525	0.485
	mAP (macro)	0.152	0.152
test (13,710 samples)	macro-F1	0.137	0.149
	micro-F1	0.508	0.433
	mAP (macro)	0.150	0.150

Per-class thresholds are shipped as thresholds.json in this repo. Use them for best F1; use sigmoid > 0.3 (or 0.15 for higher recall) as a simple default. mAP is threshold-independent and the honest number for downstream ranking-quality use cases (top-k, similarity search).

Class coverage: 2,182 species in vocab; 732 have positives in the val split (long-tail — many rare species have zero test-set representation). Median calibrated threshold: 0.30. Range: 0.02–0.94.

Full report at eval_report.json in this repo. Reproduce with:

python -m soundscape.calibrate_and_eval \
  --config soundscape/configs/ast_nonbird.yaml \
  --checkpoint model.safetensors --arch ast \
  --out-dir report/

vs BirdNET on non-bird taxa (release headline)

Head-to-head on 300 randomly-sampled non-bird test clips (insects, amphibians, mammals, reptiles). Top-1 species identification:

model	non-bird top-1 accuracy
BirdNET (v2.4, bird-focused)	6.7%
Nocturne v1.1 teacher	76.7%

That's an 11.5× lift. BirdNET catches ~7% because a handful of amphibians and mammals slip into its vocab, but for the non-avian half of the soundscape it's the wrong tool and Nocturne is the fit-for-purpose one. This is the point of the release.

Taxonomic coverage (unchanged from v1)

Taxonomic class	Species in vocab	Source
Insecta	745 (iNat) + 459 (InsectSet459)	iNat 2024 non-bird subset + InsectSet459
Amphibia	650	iNat 2024
Mammalia	296	iNat 2024
Reptilia	32	iNat 2024
Total (merged)	2182	—

Intended use

• Passive acoustic monitoring for biodiversity surveys, especially where BirdNET/Perch leave gaps.
• Agricultural pest early warning (Orthoptera + Cicadidae are heavily represented via InsectSet459).
• Amphibian presence surveys.
• Ecological research and citizen-science integration.

Out of scope

• Absence detection — a negative prediction means "no confident detection," not "the species is absent."
• Bird identification — use BirdNET / Perch. Nocturne can separate bird from non-bird signal but is not tuned for species-level bird ID.
• Bat echolocation — most is > 8 kHz relative to our 16 kHz sample rate. Use a dedicated ultrasonic pipeline.
• Legal or conservation determinations without human review.

Training data

Dataset	Files	Species	License	Role
InsectSet459 (Zenodo 18554693)	26,298	459	CC-BY-4.0	Insect head fine-tune
iNat Sounds 2024 non-bird	25,983	1,723	MIT	Multi-taxon head
iNat Sounds 2024 birds	111,029	3,846	MIT	Backbone exposure only, not in output vocab

Total training exposure: ~42k annotated non-bird clips across 2182 species.

Training recipe (v1.1)

• Backbone: AudioSet-pretrained AST (86M params). 3-epoch head warm-up with backbone frozen; then joint fine-tune.
• Optimizer: AdamW — backbone 1.5e-4, head 3.0e-3 (3× v1), cosine schedule, weight decay 0.01, grad clip 1.0.
• Loss: focal BCE (γ=2.0) for long-tail multi-label imbalance.
• Augmentation: SpecAugment (2×32 freq masks, 2×40 time masks), MixUp (α=0.3).
• Sampler: √-frequency class-balanced.
• Precision: bf16 mixed precision + gradient checkpointing.
• Batch: 32.
• 40 epochs total, best-of-N by val macro-F1 tracked.

Inference (Python)

import torch, soundfile as sf, torchaudio.functional as AF
from safetensors.torch import load_file
import json
from huggingface_hub import snapshot_download

local = snapshot_download("stratus-labs/nocturne-v1-teacher")
state = load_file(f"{local}/model.safetensors")
vocab = json.loads(open(f"{local}/vocab.json").read_text())
# Then reconstruct the ASTClassifier from soundscape/model.py and load_state_dict(state, strict=False)

Or just hit the public endpoint (curl example above).

Limitations and ethical considerations

• Coverage bias. Over-represents temperate-zone Orthoptera and Cicadidae (via InsectSet459) and geographies over-represented on iNaturalist. Under-represented: tropical arthropods, deep-forest amphibians, most reptiles.
• Recording bias. iNat recordings are opportunistic. Do not interpret model confidence as population density.
• Absence != silence. Many target species have narrow calling windows (breeding season, temperature, time of day). A survey without detections is not a survey without species.
• Dual-use. Species localization from acoustic recordings can be misused (e.g. locating rare taxa for illegal collection). Consider redacting precise GPS in public datasets built with this model.
• Data licensing. Downstream users must comply with InsectSet459 (CC-BY-4.0) and iNat 2024 (MIT). Redistribution of raw training clips is not permitted; this model is a derived work.

Citation

@misc{stratuslabs2026nocturne,
  title  = {Nocturne: Bioacoustic Species Recognition for Non-Bird Taxa},
  author = {Stratus Labs},
  year   = {2026},
  publisher = {Hugging Face},
  howpublished = {\url{https://huggingface.co/stratus-labs/nocturne-v1-teacher}}
}

Please also cite the training datasets:

• InsectSet459 (Zenodo 10.5281/zenodo.18554693)
• iNat Sounds 2024 (Chasmai et al., NeurIPS 2024)