你即將離開本站

並前往https://arxiv.org/abs/1711.05101

[1711.05101] Decoupled Weight Decay Regularization - arXiv

... (when rescaled by the learning rate), but as we demonstrate this is \emph{not} the case for adaptive gradient algorithms, such as Adam.

確定！ 回上一頁

查詢 「AdamW optimizer」的人也找了：

  1. AdamW PyTorch
    2.
  2. adamw介紹
    3.
  3. AdamW paper
    4.
  4. AdamW weight decay
    5.
  5. Adam weight decay
    6.
  6. Fixing weight decay regularization in Adam
    7.
  7. Adam optimizer learning rate
    8.
  8. BERT optimizer
    9.