{"id":"espnet-owsm-ctc-v4-1b","nom":"OWSM-CTC v4","editeur":"ESPnet","pays":null,"poids_ouverts":true,"poids_selon":null,"licence":"CC BY 4.0","licence_selon":null,"liens":{"page":"https://huggingface.co/espnet/owsm_ctc_v4_1B","essayer":"https://huggingface.co/spaces/espnet/owsm-ctc-v4"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/espnet-owsm-ctc-v4-1b/","releve":"2026-10-19","en_bref":"42e sur 53 en transcription en anglais (Open ASR).","presentation":"Modèle ouvert de reconnaissance vocale multilingue, de type encodeur seul avec décodage CTC non autorégressif, qui fait aussi de la traduction de la parole et de l'identification de la langue. Il reproduit l'entraînement de type Whisper sur 320 000 heures de données audio publiques, avec code, poids et journaux d'entraînement publiés.","sortie":"2025-05-31","taille":"1,01 milliard de paramètres","langues":{"annoncees":{"annoncees":"multilingue, nombre de langues non annoncé"},"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["langues","presentation"],"lu":"https://huggingface.co/espnet/owsm_ctc_v4_1B/raw/main/README.md (« is a novel encoder-only speech foundation model based on hierarchical multi-task self-conditioned CTC. It supports multilingual speech recognition, speech translation, and language identification within a single non-autoregressive model. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/espnet/owsm_ctc_v4_1B/raw/main/README.md (« is trained for three epochs on 320k hours of public audio data covering multilingual speech recognition, any-to-any speech translation, and language identification. »), 17/10/2026"},{"appuie":["presentation"],"lu":"https://huggingface.co/espnet/owsm_ctc_v4_1B/raw/main/README.md (« The code, pre-trained model weights, and training logs are publicly released to promote open science in speech foundation models. »), 17/10/2026"},{"appuie":["taille"],"lu":"https://huggingface.co/espnet/owsm_ctc_v4_1B/raw/main/README.md (« | OWSM-CTC v4 medium | 1.01B | https://huggingface.co/espnet/owsm_ctc_v4_1B | »), 17/10/2026"},{"appuie":["sortie"],"lu":"https://arxiv.org/html/2506.00338v1 (« arXiv:2506.00338v1 [cs.CL] 31 May 2025 »), 17/10/2026"}],"classements":[{"id":"transcription/open-asr-anglais","nom":"Transcription en anglais, Open ASR","rang":42,"sur":53,"valeur":8.08,"unite":"%","chiffres":{"rtfx":774.79,"jeux":{"AMI-Cleaned":11.1,"Earnings22-Cleaned-AA-chunked":10.31,"Gigaspeech-Cleaned":9.29,"LS Clean":2.11,"LS Other":4.38,"Private (conversational)":16.1,"Private (scripted)":6.41,"SPGISpeech":2.31,"URGENT24":16.12,"Voice Arena Monsoon":6.66,"Voxpopuli-AA-Cleaned":4.09}},"voisins":["whisper-large-v3-turbo","meta-omniasr-llm-7b-v2"],"source":{"nom":"Open ASR Leaderboard","mention":"Source : Open ASR Leaderboard, Hugging Face","url":"https://huggingface.co/spaces/hf-audio/open_asr_leaderboard","releve":"2026-10-19"}}],"noms_dans_les_sources":{"hf_audio":"espnet/owsm_ctc_v4_1B"}}