KaeLiuChenyu/ntu_diar

NTU Diarization

---

Folders for ntu_diar

scr

Main script:

1. scr
   1. dataio
   2. model
   3. task
   4. utils

i.dataio

Includes dataset processing scripts for creating Datasets and Dataloader.

ii.model

The components that make up the Diarization model. This includes feature extraction network, data augmentation, clustering, etc.

iii.task

Includes scripts to implement tasks. Including training task and inference task.

iv.utils

Utilities

task

Training & Inference task:

1. task
   1. Inference
   2. Train

i.Inference

Includes several Inference scripts for diarization models.

ii.Train

Includes several Training scripts for diarization models.

module

Module file:

1. module
   1. pre_train
   2. train

i.pre_train

Includes pre-trained model files

ii.Train

Includes checkpoint files and log files during training.

additional

The attached scripts, such as the preparation of data sets:

Usage

Train X-vector

Please check ntu_diar/task/Train/x_vector

Inference

Please check ntu_diar/task/inference/silero_xvector