Kotoba‑Whisper v2.0 is a Japanese ASR model distilled from OpenAI Whisper large‑v3, trained on 7.2 million ReazonSpeech clips, that runs 6.3× faster while matching the teacher’s CER/WER on in‑domain tests and includes stable‑ts/punctuation support and full training code on GitHub.
Description in other languages
Statistics
Statistics as of Nov 16, 2026
Hugging Face downloads (last 30 days)
25,000
Hugging Face likes
103
Listed
Sep 23, 2024
Downloads (30-day window) over time
Show as a table
| Month | DL (30d) |
|---|---|
| Dec 2025 | 124,000 |
| Jan 2026 | 3,000 |
| Feb 2026 | 4,000 |
| Mar 2026 | 5,000 |
| Apr 2026 | 5,000 |
| May 2026 | 12,000 |
| Jun 2026 | 19,000 |
| Jul 2026 | 8,000 |
| Aug 2026 | 13,000 |
| Sep 2026 | 22,000 |
| Oct 2026 | 22,000 |
| Nov 2026 | 25,000 |
Downloads over the 30 days before the last snapshot of each month.
Popular in the same category
- wav2vec2-large-xlsr-53-japanesejonatasgrosman16MJapanese wav2vec‑2 XLSR‑53 fine‑tuned on Common Voice 6.1, CSS10, and JSUT, requiring 16 kHz audio and usable via HuggingSound or HuggingFace pipelines.
- wav2vec2-large-xlsr-japanese-hiraganavumichien153kA Japanese speech‑recognition model fine‑tuned from facebook/wav2vec2‑large‑xlsr‑53 on Common Voice and the JSUT corpus, optimized for 16 kHz audio input.
- anime-whisperlitagin33kAnime Whisper is a lightweight Japanese ASR model fine‑tuned on ~5,300 h of anime‑style dialogue that delivers low hallucination, rhythm‑aligned punctuation and accurate transcription of non‑verbal sounds and NSFW content, and must be run without an initial prompt.
- kotoba-whisper-v2.2kotoba-tech16kKotoba‑Whisper‑v2.2 is a Japanese ASR model extending kotoba‑whisper‑v2.0 with integrated diarization and automatic punctuation via a HuggingFace‑Transformers pipeline, built in collaboration with Asahi Ushio and Kotoba Technologies.
- Qwen3-ASR-1.7B-JAneosophie13kFine‑tuned Qwen3‑ASR‑1.7B for Japanese ASR, optimized to accurately transcribe proper nouns, organization and product names, kanji‑heavy and mixed Japanese/English technical terms.
- kotoba-whisper-v2.0-fasterkotoba-tech13kKotoba Whisper v2.0 is converted to CTranslate2 format for use with CTranslate2 and faster‑whisper, providing installation, inference examples, Apple M2 benchmarking, and conversion instructions.
Resources in similar research fields
- japanese-gpt-neox-smallrinna656kA 12‑layer, 768‑hidden Japanese GPT‑NeoX model trained on CC‑100, C4, and Wikipedia, compatible with Huggingface, with an optional toy prefix‑tuning weight that forces each sentence to end with a smiling‑face emoji.Shared: Language Models・Semantic Text Processing
- llm-jp-3-150mllm-jp75kLLM‑jp‑3‑150m—a 150‑M‑parameter Japanese language model from the National Institute of Informatics’ LLM R&D Center—is distributed in Hugging Face Transformers format, requiring torch ≥ 2.3.0, transformers ≥ 4.40.1, accelerate ≥ 0.29.3, flash‑attn ≥ 2.5.8, and was pretrained on Japanese Wikipedia, Common Crawl, WARP/PDF, WARP/HTML, and Kaken data using a unigram byte‑fallback tokenizer.Shared: Language Models・Semantic Text Processing
- sentence-bert-base-ja-mean-tokens-v2sonoisa53kA Japanese Sentence‑BERT v2, fine‑tuned on cl‑tohoku/bert‑base‑japanese‑whole‑word‑masking with MultipleNegativesRankingLoss, boosts accuracy by ~1.5–2 % over v1 and is released as sonoisa/sentence‑bert‑base‑ja‑mean‑tokens‑v2.Shared: Language Models・Semantic Text Processing
- modernbert-ja-310msbintuitions36kModernBERT‑Ja‑310M is a Japanese‑language BERT variant that blends local‑global attention with RoPE, trained on 4.09 T tokens of Japanese/English text, supports a 102 400‑word vocabulary, 8 192‑token sequences, and is optimized for Flash Attention 2.Shared: Language Models・Semantic Text Processing
- deberta-v2-base-japaneseku-nlp24kJapanese DeBERTa V2 base model pretrained on 171 GB of Japanese Wikipedia, CC‑100 and OSCAR data using Juman++ segmentation and SentencePiece tokenization, trained for three weeks on eight NVIDIA A100 GPUs and ready for fine‑tuning.Shared: Language Models・Semantic Text Processing
- transformers-ud-japanese-electra-base-ginza-510megagonlabs22kja_ginza_electra is a spaCy v3 Python package offering a Japanese ELECTRA model fine‑tuned on mC4 and UD_Japanese_BCCWJ r2.8 (based on megagonlabs/transformers‑ud‑japanese‑electra‑base‑discrimininator) with custom bunsetu‑phrase detection, distributed under the MIT license.Shared: Language Models・Semantic Text Processing