Whisper large-v3

OpenAI

Nije hostano — mock backend Otvoreni utezi

Suverenost i hosting

Nije hostano — mock backend

Nije hostano na LLM EU u ovom postavljanju: na zahtjeve odgovara ugrađeni mock backend u regiji dev-mock. Ne izvršavaju se nikakve težine.

Sažetak

Whisper large-v3 is OpenAI’s multilingual speech-recognition and speech-translation model with 1.54B parameters under the Apache-2.0 licence. It has a fixed 30-second receptive field and processes longer audio with a sliding window, and it uses 128 mel frequency bins rather than the 80 of earlier versions. The model card warns that it was evaluated on roughly ten languages and should be evaluated in-domain before deployment.

Specifikacije

Pružatelj
OpenAI (US)
Kontekstni prozor
Nije provjereno: pružatelj nije dokumentirao ovu tvrdnju.
Parametri
1.54B
Modaliteti
audio
Licenca
apache-2.0 · licence
Otvoreni utezi
yes
Datum izlaska
07. stu 2023.
Zadaci
translation, multilingual-support, summarization

Jezici

Jako: en, de, fr, es, it, pt, nl, pl

Zadovoljavajuće: sv, da, fi, cs, ro, el, uk, ru, tr, ar, zh, ja, ko, hi

Jezici

Jako

  • en
  • de
  • fr
  • es
  • it
  • pt
  • nl
  • pl

Zadovoljavajuće

  • sv
  • da
  • fi
  • cs
  • ro
  • el
  • uk
  • ru
  • tr
  • ar
  • zh
  • ja
  • ko
  • hi

Nije provjereno: pružatelj nije dokumentirao ovu tvrdnju. — a language is listed as strong only where a source claims it; the card links the evidence where one exists.

Kada ga koristiti

Prednosti

  • one model for transcription and translation to English
  • long-standing tooling support across every major inference stack
  • Apache-2.0 licence

Ograničenja

  • fixed 30-second receptive field; longer audio needs chunking logic
  • the card only claims strong recognition for about ten languages
  • the card advises against using it for classification or speaker attributes