tango4j / auto-tuning-spectral-clustering Goto Github PK

Thank you for your open source codes.
In the paper, you used k-means clustering with the estimated number of speakers to obtain the final results.
But in the code sklearn's spectral clustering is used.
https://github.com/tango4j/Auto-Tuning-Spectral-Clustering/blob/master/spectral_opt.py#L608
Could you tell me why you changed this part?
Thanks.

hello:
Hello, I encountered a problem when running this code. Here is the situation:
(env_nmesc) root@9cbf99057050:/data/Auto-Tuning-Spectral-Clustering# ./run_demo_clustering.sh
=== [INFO] The python_envfolder exists: /data/Auto-Tuning-Spectral-Clustering/env_nmesc
/data/Auto-Tuning-Spectral-Clustering/env_nmesc/bin/python
/data/Auto-Tuning-Spectral-Clustering/sc_utils /data/Auto-Tuning-Spectral-Clustering
=== [INFO] Starting Script: affinity_score.py
./score_embedding.sh --cmd run.pl --mem 5G --python_env /data/Auto-Tuning-Spectral-Clustering/env_nmesc/bin/activate --nj 1 --score_ metric cos --out_dir /data/Auto-Tuning-Spectral-Clustering/sample_CH_xvector/cos_scores /data/Auto-Tuning-Spectral-Clustering/sample _CH_xvector/xvector_embeddings /data/Auto-Tuning-Spectral-Clustering/sample_CH_xvector/cos_scores
./score_embedding.sh: line 28: parse_options.sh: No such file or directory

How to solve it?