AutoSpeech: Neural Architecture Search for Speaker Recognition

Code for this paper AutoSpeech: Neural Architecture Search for Speaker Recognition

Shaojin Ding*, Tianlong Chen*, Xinyu Gong, Weiwei Zha, Zhangyang Wang

Overview

Speaker recognition systems based on Convolutional Neural Networks (CNNs) are often built with off-the-shelf backbones such as VGG-Net or ResNet. However, these backbones were originally proposed for image classification, and therefore may not be naturally fit for speaker recognition. Due to the prohibitive complexity of manually exploring the design space, we propose the first neural architecture search approach approach for the speaker recognition tasks, named as AutoSpeech. Our evaluation results on VoxCeleb1 demonstrate that the derived CNN architectures from the proposed approach significantly outperform current speaker recognition systems based on VGG-M, ResNet-18, and ResNet-34 back-bones, while enjoying lower model complexity.

Results

Our proposed approach outperforms speaker recognition systems based on VGG-M, ResNet-18, and ResNet-34 backbones. The detailed comparison can be found in our paper.

Method	Top-1	EER	Parameters	Pretrained model
VGG-M	80.50	10.20	67M	iden/veri
ResNet-18	79.48	12.30	12M	iden, veri
ResNet-34	81.34	11.99	22M	iden, veri
Proposed	87.66	8.95	18M	iden, veri

Visualization

left: normal cell. right: reduction cell

Quick start

Requirements

• Python 3.7

• Pytorch>=1.0: pip install torch torchvision

• Other dependencies: pip install -r requirements

Dataset

VoxCeleb1: You will need DevA-DevD and Test parts. Additionally, you will need original files: vox1_meta.csv, iden_split.txt, and veri_test.txt from official website. Alternatively, the dataset can be downloaded using dl_script.sh.

The data should be organized as:

• VoxCeleb1
  • dev/wav/...
  • test/wav/...
  • vox1_meta.csv
  • iden_split.txt
  • veri_test.txt

Running the code

• data preprocess:

  python data_preprocess.py /path/to/VoxCeleb1

  The output folder of it should be:

  • feature
    • dev
    • test
    • merged

  dev and test are used for verification, and merged are used for identification.

• Training and evaluating ResNet-18, ResNet-34 baselines:

  python train_baseline_identification.py --cfg exps/baseline/resnet18_iden.yaml

  python train_baseline_verification.py --cfg exps/baseline/resnet18_veri.yaml

  python train_baseline_identification.py --cfg exps/baseline/resnet34_iden.yaml

  python train_baseline_verification.py --cfg exps/baseline/resnet34_veri.yaml

  You need to modify the DATA_DIR field in .yaml file.

• Architecture search:

  python search.py --cfg exps/search.yaml

  You need to modify the DATA_DIR field in .yaml file.

• Training from scratch for identification:

  python train_identification.py --cfg exps/scratch/scratch.yaml --text_arch GENOTYPE

  You need to modify the DATA_DIR field in .yaml file.

  GENOTYPE is the search architecture object. For example, the GENOTYPE of the architecture report in the paper is:

  "Genotype(normal=[('dil_conv_5x5', 1), ('dil_conv_3x3', 0), ('dil_conv_5x5', 0), ('sep_conv_3x3', 1), ('sep_conv_3x3', 1), ('sep_conv_3x3', 2), ('dil_conv_3x3', 2), ('max_pool_3x3', 1)], normal_concat=range(2, 6), reduce=[('max_pool_3x3', 1), ('max_pool_3x3', 0), ('dil_conv_5x5', 2), ('max_pool_3x3', 1), ('dil_conv_5x5', 3), ('dil_conv_3x3', 2), ('dil_conv_5x5', 4), ('dil_conv_5x5', 2)], reduce_concat=range(2, 6))"

• Training from scratch for verification:

  python train_verification.py --cfg exps/scratch/scratch.yaml --text_arch GENOTYPE

• Evaluation:

  • Identification

    python evaluate_identification.py --cfg exps/scratch/scratch_iden.yaml --load_path /path/to/the/trained/model

  • Verification

    python evaluate_verification.py --cfg exps/scratch/scratch_veri.yaml --load_path /path/to/the/trained/model

Citation

If you use this code for your research, please cite our paper.

@misc{ding2020autospeech,
    title={AutoSpeech: Neural Architecture Search for Speaker Recognition},
    author={Shaojin Ding and Tianlong Chen and Xinyu Gong and Weiwei Zha and Zhangyang Wang},
    year={2020},
    eprint={2005.03215},
    archivePrefix={arXiv},
    primaryClass={eess.AS}
}

Acknowledgement

Part of the codes are adapted from deep-speaker and Real-Time-Voice-Cloning.