Automatic Speech Recognition on LibriSpeech 100h (test-other)

4.6Word Error Rate

SPEARs+a XLarge

Updated 1mo ago

Evaluation Results

Method	Links
SPEARs+a XLarge 2025.10		4.6
SPEARs Large 2025.10		4.7
SPEARs+a Large 2025.10		4.8
WavLM (Continuous) 2026.06		5.5
SPEARs Base 2025.10		5.7
HuBERT (Continuous) 2026.06		5.7
SPEARs+a Base 2025.10		6
WavLM Large 2025.10		6.1
Posterior-based soft assignment (WavLM) 2026.06		6.3
WavLM (Soft Training) 2026.06		6.6
WavLM (Hard Assignment) 2026.06		6.6
Posterior-based soft assignment (HuBERT) 2026.06		6.8
ASR-only 2026.01		7
HuBERT (Hard Assignment) 2026.06		7
Discrete WavLM 2026.01		7.1
Posterior-based soft assignment (WavLM) 2026.06		7.1
HuBERT (Soft Training) 2026.06		7.2
Posterior-based soft assignment (HuBERT) 2026.06		7.3
WavLM (Hard Assignment) 2026.06		7.4
USAD Large 2025.10		7.7
HuBERT (Hard Assignment) 2026.06		7.7
WavLM Base + 2025.10		8.4
Phonological Tokenizer 2026.01		8.5
Whisper-M + Llama 8B (4bit, r=4) 2026.06		9.2
Whisper-M + Llama 8B (4bit, r=4) 2026.06		9.3
Whisper-M + Llama 8B (4bit, r=4) 2026.06		9.4
Posterior-based soft assignment (WavLM) 2026.06		10.2
Whisper-M + Llama 8B (4bit, r=4) 2026.06		10.5
Posterior-based soft assignment (HuBERT) 2026.06		11.1
WavLM (Hard Assignment) 2026.06		11.3
USAD Base 2025.10		11.4
HuBERT (Hard Assignment) 2026.06		12.2
Whisper-M 2026.06		12.2
SPEARa Large 2025.10		18.6
Branchformer + GFSA 2023.12		22.25
Branchformer 2023.12		22.43
SPEARa Base 2025.10		23
SpeechTokenizer 2026.01		23.5
Transformer + GFSA 2023.12		24.3
Transformer 2023.12		25.42
Voc-only 2026.01		27.7
WavTokenizer 2026.01		96.8