J-HARD-TTS-Eval is a benchmark dataset for evaluating autoregressive Japanese Text-To-Speech models. It focuses on specific failure modes including stability in short sequences, repetition handling, and context completion. The dataset was created by Parakeet-Inc and last updated in January 2026.
Use Cases
- Benchmark TTS model robustness on failure modes like short sequence stability using Japanese audio and text data
- Evaluate model performance on repetition handling tasks within Japanese speech synthesis
- Assess context completion capabilities of autoregressive TTS models for Japanese language
Strengths
- Benchmark designed for specific TTS failure modes: short sequences, repetition, and context completion
- Dataset is optimized for Parquet format and includes tags for Japanese language and audio modality
- Last updated in January 2026, indicating recent maintenance
Limitations
- Unknown row count, column names, and file size limit analytical planning
- Sample data unavailable prevents preview of data structure and content
- Specific geographic coverage within Japan is unspecified
Provenance
- Source
- Parakeet-Inc via Hugging Face
- Collection Method
- Benchmark dataset for evaluating TTS model robustness
- Time Range
- null
- Freshness
- 2026-01-28
- Geography
- Japan (Language: ja)