Code

Pin 2024 03 27#15

Code

Merged

zjjott merged 31 commits intomasterintelligent-machine-learning/torch_xla:masterfrom pin_2024_03_27intelligent-machine-learning/torch_xla:pin_2024_03_27Copy head branch name to clipboardMar 27, 2024

+1,652-358

Commits

Commits on Jan 8, 2024

• Add workflow file for TPU CI (pytorch#6198)

  mbzomowski

  authored
  5801f7e

• remove model training using openxla backend (pytorch#6262)

  zpcore

  authored
  a89ebef

• test_zero1 passes on GPU with latest torch/xla (commit e1c94df) (pytorch#6268)

  Show description for a60f8e7
  jeffhataws

  authored
  a60f8e7

• Set min TPU nodes to 0 for TPU CI & disable workflow run on push (pytorch#6270)

  mbzomowski

  authored
  257f0f5

Commits on Jan 9, 2024

• Measure GPU wall time without launch overheads. (pytorch#6233)

  Show description for bd82beb
  golechwierowicz

  authored
  bd82beb

• Add matmul microbenchmark. (pytorch#6259)

  Show description for 88d97cc
  golechwierowicz

  authored
  88d97cc

• Revert "remove model training using openxla backend (pytorch#6262)" (pytorch#6276)

  zpcore

  authored
  551c0df

• Fix a typo in aten_xla_type.cpp (pytorch#6271)

  bhavya01

  authored
  f2416cf

• Revert "Remove reshape lowering (pytorch#6230)" (pytorch#6280)

  Show description for c3a5909
  cota

  authored
  c3a5909

• Limit number of bazel jobs in the build (pytorch#6279)

  will-cromar

  authored
  9a4ef68

Commits on Jan 10, 2024

• Adapt a few torch.utils.checkpoint functions for PyTorch/XLA. (pytorch#6178)

  ysiraichi

  authored
  ebb200b

• Remove or simplify hardcoded lists of device types (pytorch#6235)

  will-cromar

  authored
  050a240

• Backport PjRtStreamExecutorLoadedExecutable:GetCompileOptions (pytorch#6282)

  will-cromar

  authored
  235b82b

• Temporarily disable GPU coverage workflow in CI (pytorch#6285)

  yeounoh

  authored
  83b7571

• Add CPU PJRT plugin for testing (pytorch#6253)

  will-cromar

  authored
  2f1334d

• Update example Kaggle notebook to PyTorch/XLA 2.1 (pytorch#6287)

  will-cromar

  authored
  0e735de

Commits on Jan 11, 2024

• Update configuration.yaml to include XLA_DISABLE_FUNCTIONALIZATION (pytorch#6290)

  wonjoolee95

  authored
  4b1eaee

• Fix some more core aten opset tests (pytorch#6284)

  wonjoolee95

  authored
  9f1afbd

• Add single graph option as a runnable script. (pytorch#6277)

  golechwierowicz

  authored
  ddf4be2

• Fix output type of custom calls while lowering quant/dequant torch op to HLO (pytorch#6283)

  sdasgup3

  authored
  68f4750

• Set XLA_USE_SPMD for spmd cpp tests. (pytorch#6273)

  vanbasten23

  authored
  a728afe

• Ignore non-XLA nodes and their direct dependents. (pytorch#6170)

  ysiraichi

  authored
  8141078

• [Core Aten] Add and enable tests for aten index_select and logical_and (pytorch#6293)

  chunnienc

  authored
  04e1238

• Support of implicit broadcasting with unbounded dynamism (pytorch#6219)

  sdasgup3

  authored
  896de17

Commits on Jan 12, 2024

• Track mean relative error across models. (pytorch#6250)

  Show description for 30c6b01
  golechwierowicz

  authored
  30c6b01

• Enable Tensor index check in both dimension (pytorch#6299)

  ManfeiBai

  authored
  2850288

Commits on Jan 13, 2024

• Correct order of constant vs args (pytorch#6300)

  qihqi

  authored
  5517779

• Fix some more aten core ops (pytorch#6301)

  wonjoolee95

  authored
  5d0de94

• fix gpu spmd core dump (pytorch#6275)

  mars1248

  authored
  de27258

Commits on Jan 14, 2024

• Fix global_runtime_device_count() for spmd case (pytorch#6266)

  vanbasten23

  authored
  27ae6dc

Commits on Mar 27, 2024

• Merge commit '27ae6dc955734b223e2b93e1c5bbcd681b0537fd'

  

  mars1248

  authored and

  root

  committed
  0afee95