1. 16 Aug, 2019 2 commits
    • Deyu Fu's avatar
      clean up variance options support by all fused optimizers: · 18062b69
      Deyu Fu authored
      correctly not apply bias correction to epsilon(same as recent upstream change)
      correctly not apply bias correction to weight decay(consistent with upstream AdamW)
      Make adam_w_mode for FusedAdam/LAMB, to do L2 or Weight Decay (Adam vs AdamW)
      Correct document reg_inside_moment differently from adam_w_mode in FusedNovoGrad
      Removed legacy eps_mode from FusedAdam
      Make internal math type float across fused optimizers
      18062b69
    • Deyu Fu's avatar
      add fused lamb, put lamb kernels into one file · c8f9cceb
      Deyu Fu authored
      c8f9cceb