WHISTRESS extends Whisper with a token-level stress classifier trained on a new synthetic dataset, and shows zero-shot transfer to natural speech benchmarks.
Specifically, we propose generating two outputs for each input speech signal: the tran- scription and an emphasis score for each token in the transcrip- tion
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
WHISTRESS: Enriching Transcriptions with Sentence Stress Detection
WHISTRESS extends Whisper with a token-level stress classifier trained on a new synthetic dataset, and shows zero-shot transfer to natural speech benchmarks.