This preprocess dataset is used for ConPCO: Preserving Phoneme Characteristics For Automatic Pronunciation Assessment Leveraging Contrastive Ordinal Regularization, which is published at IEEE ICASSP 2025.
The dataset has a total size of approximately 12 GB. To facilitate faster downloads, we have compressed the files into ZIP archives. Please unzip the archives after downloading to access the data.
The Gihub repository of ConPCO is at here.