pytorch/xla

Pull Requests Commits

fixed reloading of sharded master weights

avizon-aws committed 1 year ago

4eca7f5d

Document test/neuron/run_tests.sh. (#9205)

zhanyong-wan committed 1 year ago

Verified e28174c9

Improve CI build speed by letting bazel decide how many worker threads to have (#9209)

zhanyong-wan committed 1 year ago

Verified eedc8d23

Fix the examples in API_GUIDE (#9213)

zhanyong-wan committed 1 year ago

Verified 7ce8493d

Reflow ci.md to 80 chars (#9151)

tengyifei committed 1 year ago

Verified 334b4d3f

Update ci.md and fix typo (#9150)

tengyifei committed 1 year ago

Verified 0b6349b0

Clean up env var usages in run_tests.sh scripts. (#9207)

zhanyong-wan committed 1 year ago

Verified 6e3f4076

Split expensive TPU tests to run in parallel (#9198)

ghpvnist committed 1 year ago

Verified b5937198

Make 2 more run_tests.sh support test selection (#9192)

zhanyong-wan committed 1 year ago

Verified 01b5408d

Allow run_tests to run a subset of the tests. (#9190)

zhanyong-wan committed 1 year ago

Verified edc1a882

[Kernel] Use call_jax to simplify the gmm pallas kernel wrapper (#9180)

yaochengji committed 1 year ago

Verified f39434ab

get master ip address for neuron device (#9120)

aws-zhenguo committed 1 year ago

Verified fd10b960

test: add test_xla_graph_execution to test flags (_set_allow_execution with PT_XLA_DEBUG_LEVEL) (#9171)

aws-yyjau committed 1 year ago

Verified 3781920c

Replace `xm.mark_step` with `torch_xla.sync()` wherever possible (#9070)

ghpvnist committed 1 year ago

Verified 8c71b049

Add assume_pure_torch implementation for forward pass only (#9135)

bhavya01 committed 1 year ago

Verified e724225f

fix: reload the param groups in optimizer (#9164)

avizon-aws committed 1 year ago

Verified b8c42523

Replace implementation of `xla_torch.sync()` to `xm.mark_step()` (#9086)

ghpvnist committed 1 year ago

Verified bbafe357

Switch bazel default spawn strategy to `sandboxed` (#9168)

zhanyong-wan committed 1 year ago

Verified a9ba21cf

Fix a clang compilation error. (#9156)

zhanyong-wan committed 1 year ago

Verified ffd4488c

add aarch64 platform build support (#8663)

snadampal committed 1 year ago

Verified a061a704

Don't install wheels when building docs (#9153)

ghpvnist committed 1 year ago

Verified 05ac08f3

Update XLA to a 2025/5/13 revision. (#9155)

zhanyong-wan committed 1 year ago

Verified 1048005b

Add instructions on how to rebase local commits (#9162)

zhanyong-wan committed 1 year ago

Verified 80f64710

Add a linter check to prevent *.cc file names. (#9147)

zhanyong-wan committed 1 year ago

Verified 40810f01

Rename *.cc files to *.cpp for consistency. (#9149)

zhanyong-wan committed 1 year ago

Verified d9d8f38c

Update instructions for updating forked pytorch (#9154)

zhanyong-wan committed 1 year ago

Verified 1f9dd8f3

Re-land: Add `clang` link to PATH. (#9144)

ysiraichi committed 1 year ago

Verified dfac6042

Ensure that all of PyTorch/XLA C++ is compiled with exceptions enabled. (#9146)

zhanyong-wan committed 1 year ago

Verified 8450607a

Test torchax on Python 3.10 - 3.12 (#9139)

tengyifei committed 1 year ago

Verified 259b438e

Silence distributed warning (#9140)

tengyifei committed 1 year ago

Verified 99e2d3c9

Older