UofT-EcoSystem/hfta

Issues

• Type error with lr_scheduler

  #46 opened 2 years ago by zyfhaha
  1

• [Optim] No need to broadcastablize coefficients during __init__

  #10 opened 3 years ago by wangshangsam
  1

• [Ops] [Optim] Update to 1.9.1.

  #37 opened 3 years ago by wangshangsam
  0

• why use torch.bmm as a packing op for matrix multiplication?

  #36 opened 3 years ago by huangjundashuaige
  2

• [MLSys21] Create `mlsys21` branch after the artifact docs & tutorial are merged in.

  #31 opened 4 years ago by wangshangsam
  1

• [Doc] MLSys'21 Artifact Guide for V100 & RTX6000 experiments.

  #21 opened 4 years ago by wangshangsam
  0

• [Docker] Upload Docker Images for Reproducing MLSys'21 Artifact to DockerHub

  #24 opened 4 years ago by wangshangsam
  9

• [Docs] `Nsight Compute` -> `Nsight System` in docs/mlsys21/a100_throughputs.md

  #29 opened 4 years ago by wangshangsam
  1

• [Docs] Only `nvidia20.06` version docker image should be built for A100

  #30 opened 4 years ago by wangshangsam
  1

• [Tutorial] Add "Getting Started" Colab Tutorial

  #20 opened 4 years ago by wangshangsam
  0

• [Docs] Create HFTA's home website

  #32 opened 4 years ago by wangshangsam
  0

• [Test] Setup automation for unit tests & integration/regression tests.

  #26 opened 4 years ago by wangshangsam
  0

• [Test] Setup integration & regression tests.

  #25 opened 4 years ago by wangshangsam
  0

• [Install] Check Installation Guide with TPU v3

  #22 opened 4 years ago by wangshangsam
  1

• [Upgrade] Check HFTA's compatibility for PyTorch > 1.7

  #19 opened 4 years ago by wangshangsam
  0

• [PyPi] Upload PyPi package.

  #23 opened 4 years ago by wangshangsam
  0

• [workflow] Plan B reliability issue with Transformer on TPU

  #15 opened 4 years ago by xinli-git
  15

• [Optim] Add unittests for PyTorch native AMP support

  #11 opened 4 years ago by wangshangsam
  0

• Unused `${modes_flag}` in `_workflow_pointnet.sh`

  #9 opened 4 years ago by wangshangsam
  2