Skip to content

Releases: KakaruHayate/ColorSplitter

1.0.1

13 Jan 15:46
Compare
Choose a tag to compare

New version features

Implemented automatic optimization of clustering results, no longer need users to judge the optimal clustering results themselves.

splitter.py deleted the --nmax parameter, added --nmin (minimum number of timbre types, invalid when cluster parameter is 2) --cluster (clustering method, 1:SpectralCluster, 2:UmapHdbscan), --mer_cosine to merge clusters that are too similar.

New version tips

  1. Run splitter.py directly with the default parameters by specifying the speaker.

  2. If the result has only one cluster, observe the distribution map, set --nmin to the number you think is reasonable, and run splitter.py again.

  3. The optimal value of --nmin may be smaller than expected in actual tests.

  4. The new clustering algorithm is faster, it is recommended to try multiple times.

新版本特性

实装了聚类结果自动优化,不再需要用户自己判断聚类最优结果

splitter.py删除了--nmax参数,添加了--nmin(最小音色类型数量,cluster参数为2时无效)--cluster(聚类方式,1:SpectralCluster, 2:UmapHdbscan),--mer_cosine合并过于相似的簇

新版本使用技巧

1.默认参数直接指定说话人运行splitter.py

2.如果结果只有一个簇,观察分布图,将--nmin设为你认为合理的数量,再次运行splitter.py

3.实际测试下--nmin的最优值可能比想象的要小

4.新的聚类算法速度较快,建议多次尝试