This is the implementation of the following paper:

• Singh, Prachi, et. al. (2023)."Supervised Hierarchical Clustering using Graph Neural Networks for Speaker Diarization." Proceedings of ICASSP 2023. (paper)

• Prerequisites
• Installation
• Running the recipes

• Cite
• Contact

The following packages are required to run the code.

• Python >= 3.7
• Kaldi
• dscore
• Voxconverse
• AMI

The following pretrained models are provided.

• ETDNN x-vector model.
• PLDA models for Voxconverse and AMI dataset.
• SHARC models for Voxconverse and AMI.

• clone the repo:

$ git clone [email protected]:prachiisc/SHARC.git
$ cd SHARC

• Create the environment: We recommend running the recipes from a fresh virtual environment. Make sure to activate the environment before proceeding.

$ conda create --name SHARC --file requirements.txt
$ conda activate SHARC

• Install Kaldi. If you are a Kaldi novice, please consult the following for additional documentation:
  • Kaldi tutorial
  • Kaldi for Dummies tutorial
• Go to cloned repository and copy kaldi path in path.sh given as:

$ local_dir="Full_path_of_cloned_repository"
$ Add "export KALDI_ROOT=/path_of_kaldi_directory/kaldi" in the first line of $local_dir/path.sh

• Create Softlinks of necessary directories:

$ local_dir="Full_path_of_cloned_repository"
$ cd $local_dir
$ . ./path.sh
$ ln -sf $KALDI_ROOT/egs/wsj/s5/utils .  # utils dir
$ ln -sf $KALDI_ROOT/egs/wsj/s5/steps .  # steps dir

• Check the data directories in tools_diar/data Change tools_diar/data/datasetname/wav.scp with your path of wavfiles.

We include full recipes for reproducing the results for Voxconverse and AMI dataset:

   bash services/test_xvec_preprocess.sh <vox_set> nj

<vox_set> : vox_diar/vox_diar_test

nj : number of jobs [min(40,number of processors available)]

   bash scripts/test_xvec_parallel.sh Vox

   bash services/test_xvec_preprocess.sh <ami_set> nj

<ami_set> : ami_dev/ami_eval

nj : number of jobs [min(15,number of processors available)]

   bash scripts/test_xvec_parallel.sh AMI

   bash services/test_xvec_preprocess.sh <train_set> <nj>

   bash scripts/train_xvec.sh <Vox/AMI>

If you are using the resource, please cite as follows:

@INPROCEEDINGS{10095372,  
  author={Singh, Prachi and Kaul, Amrit and Ganapathy, Sriram},
  booktitle={2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},  
  title={Supervised Hierarchical Clustering Using Graph Neural Networks for Speaker Diarization},  
  year={2023}, 
  volume={}, 
  pages={1-5}, 
  doi={10.1109/ICASSP49357.2023.10095372}}
  

If you have any comment or question, please contact [email protected]