RNN-WER

RNN-WER ni modeli ya usemi iliyoundwa na DeepMind na University of Toronto (Uingereza, Kanada), iliyotolewa 22 Juni 2014.

Hatujui kama inapatikana kwa umma.

Ina vigezo (parameters) milioni 26.5.

Ni moja ya modeli zinazonukuliwa sana na watafiti.

Ilipotoka, ilivunja rekodi katika kazi yake.

Naweza kuiendesha mwenyewe?

Hapana.

Chanzo chetu hakisemi kama uzito wake umetolewa. Angalia ukurasa wa walioiunda.

Modeli nyingine za DeepMind

  1. DiffusionGemma 26B-A4BDeepMindJun 2026Lugha, Utambuzi wa picha
    Wazi25.2BKompyuta yenye uwezo mkubwa
  2. Diffusion RendererNvidia, University of Illinois Urbana-Champaign (UIUC), University of Toronto na Vector InstituteMar 2025Video
    Wazi, si kwa biashara1.1BSimu
  3. Octo-BaseCarnegie Mellon University (CMU), DeepMind, Stanford University na University of California (UC) BerkeleyMei 2024Roboti
    Wazi93M
  4. Octo-SmallCarnegie Mellon University (CMU), DeepMind, Stanford University na University of California (UC) BerkeleyMei 2024Roboti
    Wazi27M
  5. Distilled GrandmasterDeepMindFeb 2024Michezo
    Wazi270M
  6. Llemma 34BCarnegie Mellon University (CMU), EleutherAI, Princeton University, University of Cambridge, University of Toronto, University of Washington na Vector InstituteOkt 2023Hisabati, Lugha
    Wazi, kwa masharti34BKompyuta yenye uwezo mkubwa