{"id":"vita-freeze-omni","nom":"Freeze-Omni","editeur":"VITA","pays":null,"poids_ouverts":true,"poids_selon":null,"licence":"Apache 2.0","licence_selon":null,"liens":{"poids":"https://huggingface.co/VITA-MLLM/Freeze-Omni","page":"https://github.com/VITA-MLLM/Freeze-Omni"},"fiche_auto":false,"page":"https://quelleia.com/voix/modeles/vita-freeze-omni/","releve":"2026-11-09","en_bref":"34e sur 38 en raisonnement à l'écoute (Big Bench Audio) ; 28e sur 31 en tenue de la conversation (Full Duplex Bench).","presentation":"Modèle de dialogue vocal à poids ouverts bâti sur un modèle de langage textuel gelé (Qwen2-7B-Instruct), pour garder ses capacités intactes. Un encodeur et un décodeur de parole en flux lui permettent de répondre vite et de détecter quand l'utilisateur l'interrompt.","sortie":"2024-11-26","langues":{"annoncees":null,"prises_en_charge_selon_open_tts":null,"evaluees":["en"]},"preuves":[{"appuie":["presentation"],"lu":"https://raw.githubusercontent.com/VITA-MLLM/Freeze-Omni/main/README.md (« Freeze-Omni is a speech-to-speech dialogue model, exhibiting the characteristic of being \"**smart**\" as it is constructed upon a \"**frozen**\" text-modality LLM. »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://raw.githubusercontent.com/VITA-MLLM/Freeze-Omni/main/README.md (« Download [Qwen2-7B-Instruct checkpoint](https://huggingface.co/Qwen/Qwen2-7B-Instruct) then move `Qwen2-7B-Instruct` into the root dir of this repo »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://raw.githubusercontent.com/VITA-MLLM/Freeze-Omni/main/README.md (« Freeze-Omni contains a speech encoder that supports streaming speech input and a speech decoder that generates streaming output speech. »), 23/10/2026"},{"appuie":["presentation"],"lu":"https://raw.githubusercontent.com/VITA-MLLM/Freeze-Omni/main/README.md (« These states will determine whether or not the user interrupts the dialogue to achieve a duplex dialogue for the user and the bot. »), 23/10/2026"},{"appuie":["sortie"],"lu":"https://raw.githubusercontent.com/VITA-MLLM/Freeze-Omni/main/README.md (« **`2024.11.26`** 🌟 The inference code, server demo and model weights **have been released**. »), 23/10/2026"}],"classements":[{"id":"voix-a-voix/raisonnement","nom":"Raisonnement à l'écoute, Big Bench Audio","rang":34,"sur":38,"valeur":33.4,"unite":"%","chiffres":{"hebergeur":"VITA"},"voisins":["openai-gpt-4o-audio","nvidia-nemotron-voicechat"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-speech","releve":"2026-11-09"}},{"id":"voix-a-voix/conversation","nom":"Tenue de la conversation, Full Duplex Bench","rang":28,"sur":31,"valeur":58.73,"unite":"%","chiffres":{"hebergeur":"VITA"},"voisins":["kyutai-moshi","nvidia-nemotron-voicechat"],"source":{"nom":"Artificial Analysis","mention":"Source : Artificial Analysis","url":"https://artificialanalysis.ai/speech-to-speech","releve":"2026-11-09"}}],"noms_dans_les_sources":{"aa_voix_a_voix":"Freeze-Omni"}}