ESPnet

ESPnet is a benchmark framework used for evaluating automatic speech recognition (ASR) tasks, including domain-shifted scenarios like child speech and regional accents.

Papers using ESPnet (3)

SSVD-O: Parameter-Efficient Fine-Tuning with Structured SVD for Speech Recognition2026SSVD: Structured SVD for Parameter-Efficient Fine-Tuning and Benchmarking under Domain Shift in ASR2025A Comparative Study on Non-Autoregressive Modelings for Speech-to-Text Generation2021 · 7 cites