espnet/interspeech2019-tutorial

Hands-on Jupyter notebooks demonstrate end-to-end neural architectures for both ASR and TTS tasks using the ESPnet framework, executable directly in Google Colab. The materials unify sequence-to-sequence modeling approaches across speech recognition and synthesis, with accompanying lecture slides covering advanced neural techniques and implementation details. Designed as a practical tutorial around the ESPnet ecosystem, it bridges the gap between theoretical methods and reproducible implementations for neural speech processing.

194 stars. No commits in the last 6 months.