kotoba-whisper-bilingual-v1.0

Home/Models/kotoba-tech/kotoba-whisper-bilingual-v1.0

kotoba-tech/kotoba-whisper-bilingual-v1.0

kind: original
quant: bf16 (16-bit)
family: misc
modality: automatic-speech-recognition
root model: —
provider: kotoba-tech
license: apache-2.0
language: en, ja
context: —
library: transformers
vRAM (GB): —
size on disk: —
downloads: 9,536
likes: 19
created: —
last modified: 2024-09-30T06:57:43+00:00 (2024)
huggingface: https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0
tags: transformerssafetensorswhisperautomatic-speech-recognitionaudiohf-asr-leaderboardenjadataset:japanese-asr/en_asr.mlsdataset:japanese-asr/ja_asr.reazon_speech_allarxiv:2212.04356arxiv:2311.00430license:apache-2.0endpoints_compatibleregion:usdeploy:azure

benchmark	category	score	conf	evidence
WER FLEURS	speech-asr	98.7	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 73.9 \| 98.7 \|`
WER AMI	speech-asr	16.7	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 16.7 \| 15.3 \| 2.4 \| 4.1 \| 8.3 \|`
WER Earnings-22	speech-asr	15.3	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 16.7 \| 15.3 \| 2.4 \| 4.1 \| 8.3 \|`
WER Common Voice	speech-asr	9.8	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 9.8 \| 9.3 \| 16.8 \|`
WER VoxPopuli	speech-asr	8.3	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 16.7 \| 15.3 \| 2.4 \| 4.1 \| 8.3 \|`
WER TED-LIUM	speech-asr	4.1	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 16.7 \| 15.3 \| 2.4 \| 4.1 \| 8.3 \|`
WER LibriSpeech	speech-asr	2.4	low	`\| [kotoba-tech/kotoba-whisper-bilingual-v1.0](https://huggingface.co/kotoba-tech/kotoba-whisper-bilingual-v1.0) \| 16.7 \| 15.3 \| 2.4 \| 4.1 \| 8.3 \|`

none scraped.