BirdNET v2.4

BirdNET v2.4 (GLOBAL 6K) - ONNX variants

ONNX builds of the BirdNET GLOBAL 6K V2.4 bird sound classifier, optimized for edge deployment in BirdNET-Go. This repo holds the precision/backend variants; the stock upstream TFLite model is unchanged and not re-hosted here.

Powered by BirdNET (https://birdnet.cornell.edu/)

BirdNET is developed by the K. Lisa Yang Center for Conservation Bioacoustics at the Cornell Lab of Ornithology and Chemnitz University of Technology. These ONNX files are derived from the upstream BirdNET v2.4 model. Attribution to BirdNET is a hard license requirement: do not strip it.

Model summary

• Classes: 6,522 species (scientific + common name, see labels.txt)
• Sample rate: 48 kHz
• Clip length: 3 s (raw PCM waveform)
• Input tensor: input, float32, shape [batch, 144000] (3 s x 48 kHz)
• Output tensor: output, float32, shape [batch, 6522] (per-class logits; apply sigmoid for confidence scores in [0, 1])

The two variants share an identical input/output interface, so they are drop-in replacements for one another.

Variants

File	Precision	Size	Backend / target	Notes
BirdNET_v2.4_int8_arm.onnx	INT8 (MatMul-only) + FP32 conv	~47 MB	ONNX Runtime on ARM / low-RAM CPU	Dynamic INT8 applied only to the 1024x6522 classification head; the CNN backbone stays FP32. ~98% top-1 agreement vs FP32. The recommended low-RAM CPU build.
BirdNET_v2.4_fp32.onnx	FP32	~62 MB	OpenVINO (and full-precision reference)	Canonical full-precision master. Under OpenVINO it runs at f16 or f32 via INFERENCE_PRECISION_HINT.

Precision notes

• CPU / ARM: use int8_arm. Full all-ops INT8 (ConvInteger) is not shipped: it breaks accuracy (~34% top-1) and has no fast ARM kernel. Only MatMul-only quantization of the head is accuracy-safe.
• OpenVINO: use fp32. The empty INFERENCE_PRECISION_HINT resolves to f16 on fp16-capable hardware (A76 NEON, AVX512-FP16) and to f32 elsewhere. Force INFERENCE_PRECISION_HINT=FP32 on GPU, where f16 miscompiles.
• f16 is intentionally not provided as a separate file: OpenVINO derives it from the FP32 master via the precision hint, and on CPU f16 uses more RAM than fp32 (the runtime up-converts f16 weights to f32 at load).

Note: this is the bird classifier. The BirdNET v2.4 backbone is also used as an embedding extractor for bat detection; that embedding model lives separately at tphakala/BattyBirdNET-onnx and must stay FP32 (its raw embedding output overflows at f16).

Labels

labels.txt has 6,522 lines, one per class, in BirdNET order. Format is Scientific name_Common name, for example:

Abroscopus albogularis_Rufous-faced Warbler

Output index i corresponds to line i of labels.txt.

Usage (ONNX Runtime, Python)

import numpy as np, onnxruntime as ort

sess = ort.InferenceSession("BirdNET_v2.4_int8_arm.onnx")

# 3 s of 48 kHz mono PCM as float32, shape [1, 144000]
audio = np.zeros((1, 144000), dtype=np.float32)

logits = sess.run(["output"], {"input": audio})[0]   # [1, 6522]
conf = 1.0 / (1.0 + np.exp(-logits))                  # sigmoid -> [0, 1]
labels = open("labels.txt").read().splitlines()
top = conf[0].argmax()
print(labels[top], float(conf[0, top]))

Checksums

See SHA256SUMS.

License

BirdNET v2.4 is distributed under CC BY-NC-SA 4.0 (non-commercial, share-alike, attribution required). See LICENSE and keep the BirdNET attribution above with any use or redistribution.

Source

• Upstream: birdnet-team/BirdNET-Analyzer
• ONNX conversion + quantization recipes: tphakala/birdnet-go