Adaptive Input Transformer + RD

Adaptive Input Transformer + RD ni modeli ya lugha iliyoundwa na Microsoft Research Asia na Soochow University (China, Taiwani), iliyotolewa 28 Juni 2021.

Haijatolewa kwa umma. Tunaijua kupitia ripoti au tangazo la walioiunda.

Ina vigezo (parameters) milioni 247.

Ilipotoka, ilivunja rekodi katika kazi yake.

Naweza kuiendesha mwenyewe?

Hapana.

Haiwezi kupakuliwa wala kutumika nje ya maabara iliyoiunda. Inaonekana hapa kwa sababu ilitangazwa au kuelezwa kwenye ripoti.

Modeli nyingine za Microsoft Research Asia

  1. Baize-v2-13B (白泽)Microsoft Research Asia, Sun Yat-sen University na University of California San DiegoDes 2023Lugha
    Wazi, si kwa biashara13BKompyuta mpakato
  2. Spec-DrafterMicrosoft Research Asia na Peking UniversityOkt 2023Lugha
    Wazi, si kwa biashara500MSimu
  3. Swin Transformer V2 (SwinV2-G)Microsoft Research AsiaNov 2021Utambuzi wa picha, Video
    Wazi3BSimu
  4. TrOCRBeihang University na Microsoft Research AsiaSep 2021Utambuzi wa picha
    Wazi558MSimu
  5. Tensorized Transformer (OBW)Beijing Institute of Technology, Microsoft Research Asia na Tianjin UniversityJun 2019Lugha
    Haijatolewa160M
  6. AWD-LSTM-MoS + dynamic evaluation (WT2, 2018)Microsoft Research Asia na Peking UniversitySep 2018Lugha
    Haijatolewa35M