HyperAI초신경
홈뉴스최신 연구 논문튜토리얼데이터셋백과사전SOTALLM 모델GPU 랭킹컨퍼런스
전체 검색
소개
HyperAI초신경

Automatic Lyrics Transcription On Jam Alt 1

평가 지표

Case-Sensitive Word Error Rate
Line break F-1
Punctuation F-1
Section break F-1
Word Error Rate (WER)

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름
Case-Sensitive Word Error Rate
Line break F-1
Punctuation F-1
Section break F-1
Word Error Rate (WER)
Paper TitleRepository
Whisper v342.571.541.42.637.7Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v3 +demucs47.266.925.8-43.0Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v3 +demucs-66.823.3-43.0Jam-ALT: A Formatting-Aware Lyrics Transcription Benchmark
AudioShake v1-80.759.077.422.1Jam-ALT: A Formatting-Aware Lyrics Transcription Benchmark
Whisper v3 +lang41.472.541.82.636.4Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v2 +demucs +lang41.353.441.8-35.6Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v2 +lang43.765.534.911.639.7Lyrics Transcription for Humans: A Readability-Aware Benchmark
OWSM v3.1 +demucs +lang69.447.321.5-63.4Lyrics Transcription for Humans: A Readability-Aware Benchmark
LyricWhiz-74.034.01.424.6Jam-ALT: A Formatting-Aware Lyrics Transcription Benchmark
Whisper v3 +demucs +lang47.266.925.8-43.0Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v2-63.031.311.243.8Jam-ALT: A Formatting-Aware Lyrics Transcription Benchmark
Whisper v2 +demucs-53.839.2-32.3Jam-ALT: A Formatting-Aware Lyrics Transcription Benchmark
LyricWhiz28.074.034.01.424.6Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v247.563.031.511.243.8Lyrics Transcription for Humans: A Readability-Aware Benchmark
AudioShake v320.984.365.384.8 17.3Lyrics Transcription for Humans: A Readability-Aware Benchmark
Whisper v3-71.540.92.637.7Jam-ALT: A Formatting-Aware Lyrics Transcription Benchmark
Whisper v2 +demucs39.153.942.2-33.3Lyrics Transcription for Humans: A Readability-Aware Benchmark
OWSM v3.1 +lang74.042.722.3-68.6Lyrics Transcription for Humans: A Readability-Aware Benchmark
0 of 18 row(s) selected.