pytorch/xla

Pull Requests Commits

Update stablehlo.py

ManfeiBai committed 1 year ago

Verified af8755f0

Handle dynamo counter change after fake tensor cache (#6308)

JackCaoG committed 1 year ago

Verified d6dc1a01

Address GPU CI coverage timeout (#6306)

yeounoh committed 1 year ago

Verified 7fae2ab1

update release build to use rc8 tag (#6310)

zpcore committed 1 year ago

Verified 1cfd0cf5

Support mixed precision for torch.where (#6303)

qihqi committed 1 year ago

Verified d49fc70d

Fix global_runtime_device_count() for spmd case (#6266)

vanbasten23 committed 1 year ago

Verified 27ae6dc9

fix gpu spmd core dump (#6275)

mars1248 committed 1 year ago

Verified de27258c

Fix some more aten core ops (#6301)

wonjoo-wj committed 1 year ago

Verified 5d0de94b

Correct order of constant vs args (#6300)

qihqi committed 1 year ago

Verified 55177798

Enable Tensor index check in both dimension (#6299)

ManfeiBai committed 1 year ago

Verified 28502884

Track mean relative error across models. (#6250)

golechwierowicz committed 1 year ago

Verified 30c6b01c

Support of implicit broadcasting with unbounded dynamism (#6219)

sdasgup3 committed 1 year ago

Verified 896de17e

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

chunnienc committed 1 year ago

Verified 04e1238a

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

ysiraichi committed 1 year ago

Verified 81410785

Set XLA_USE_SPMD for spmd cpp tests. (#6273)

vanbasten23 committed 1 year ago

Verified a728afe8

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

sdasgup3 committed 1 year ago

Verified 68f47508

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

golechwierowicz committed 1 year ago

Verified ddf4be26

Fix some more core aten opset tests (#6284)

wonjoo-wj committed 1 year ago

Verified 9f1afbd3

Update configuration.yaml to include XLA_DISABLE_FUNCTIONALIZATION (#6290)

wonjoo-wj committed 1 year ago

Verified 4b1eaeef

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

will-cromar committed 1 year ago

Verified 0e735dec

Add CPU PJRT plugin for testing (#6253)

will-cromar committed 1 year ago

Verified 2f1334d7

Temporarily disable GPU coverage workflow in CI (#6285)

yeounoh committed 1 year ago

Verified 83b7571a

Backport `PjRtStreamExecutorLoadedExecutable:GetCompileOptions` (#6282)

will-cromar committed 1 year ago

Verified 235b82b5

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

will-cromar committed 1 year ago

Verified 050a2403

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

ysiraichi committed 1 year ago

Verified ebb200b0

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

will-cromar committed 1 year ago

Verified 9a4ef68a

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

cota committed 1 year ago

Verified c3a59095

Fix a typo in aten_xla_type.cpp (#6271)

bhavya01 committed 1 year ago

Verified f2416cf8

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

zpcore committed 1 year ago

Verified 551c0dfa

Add matmul microbenchmark. (#6259)

golechwierowicz committed 1 year ago

Verified 88d97cc7

Older