Cross-lingual Transfer for Speech Processing using Acoustic Language Similarity

Indic TTS Samples can be found at https://peter-yh-wu.github.io/cross-lingual/

Code for neural ASR & TTS experiments can be found at https://github.com/peter-yh-wu/espnet

• The javanese_asr branch contains scripts to train the Javanese ASR model in egs2/java
• The sundanese_asr branch contains scripts to train the Sundanese ASR model in egs2/sunda
• The indic_tts branch contains scripts to train the Indic TTS models in egs/cmu_indic

The embs/ directory contains our acoustic-based language embeddings, where each index corresponds to the respective language in language_ids.txt. LangList.txt contains more details about each language.