Usage

import librosa
from espnet2.bin.spk_inference import Speech2Embedding

speech2embedding = Speech2Embedding.from_pretrained(model_tag="espnet/voxcelebs12_ebranchformer_base")
# the speaker models are trained on 16 kHz mono; librosa gives that from
# whatever the file holds
speech, rate = librosa.load("audio.wav", sr=16000, mono=True)
embedding = speech2embedding(speech)  # (1, embedding_dim)

RESULTS

Environments

date: 2024-11-11 13:27:22.449116

python version: 3.10.13 (main, Sep 11 2023, 13:44:35) [GCC 11.2.0]
espnet version: 202409
pytorch version: 2.2.2+rocm5.6

	Mean	Std
Target	8.1720	3.9750
Non-target	2.0805	2.0805

Model name	EER(%)	minDCF
conf/tuning/train_ebranchformer_aggregate_sys1_D384_L6_Rel_LayerDrop	0.638	0.04797

Downloads last month: 10