pytorch/xla

Pull Requests Commits

Support fp8 lowering in scalar

yaochengji committed 187 days ago

5bdbc894

Migrate runtime.xla_device in favor of core.xla_model.xla_device (#9200)

ghpvnist committed 188 days ago

Verified 022ecde6

Update OpenXLA to 2025/5/30. (#9271)

zhanyong-wan committed 188 days ago

Verified a159c9c8

Add instruction for exporting inlined constant (#8707)

qihqi committed 188 days ago

Verified 0b28030c

Remove CI jobs for GPU (#9277)

zhanyong-wan committed 188 days ago

Verified 9c21366a

Refactor GetNumDevices and create GetNumGlobalDevices (#9184)

pgmoka committed 189 days ago

Verified 1a82b709

Add scripts/update_deps.py to automate the updating the dependencies. (#9270)

zhanyong-wan committed 189 days ago

Verified fbc44603

[torchax] Support View in 'jax_view' (#9273)

lsy323 committed 191 days ago

Verified c2a9b32b

Tune CI job time-outs to be closer to the actual durations. (#9267)

zhanyong-wan committed 191 days ago

Verified cdd0202e

Update libtpu and jax to their latest nightly builds. (#9264)

zhanyong-wan committed 191 days ago

Verified 222dc4ab

In pjrt runtime client, raise a Python exception if XLA compilation fails. (#9138)

zhanyong-wan committed 192 days ago

Verified d1ca3c48

[torchax] Added support for bicubic and billinear resampling (#9222)

qihqi committed 192 days ago

Verified 2196d0cf

Run `test_collective_permute.py` on TPU CI. (#9257)

ysiraichi committed 192 days ago

Verified 06c5533d

[Kernel] support kv cache quantization in ragged attention kernel (#9249)

yaochengji committed 192 days ago

Verified 2edcd2e3

Improve `torch_xla.compile` documentation (#9194)

sdasgup3 committed 192 days ago

Verified 248a8b33

Document the difference between tracing time and execution time (#9133)

sdasgup3 committed 192 days ago

Verified 513ab7ca

Convert some XLA_CHECKs to fatal errors. (#9263)

zhanyong-wan committed 192 days ago

Verified d43142cb

Optimize build_developer.sh for the common workflow. (#9262)

zhanyong-wan committed 192 days ago

Verified 51a3bae2

Update README.md with supported Python version (#9239)

zzzwen committed 192 days ago

Verified 808773a9

[Kernel] add group_offset and transpose_rhs support in gmm kernel (#9251)

yaochengji committed 193 days ago

Verified 9871aed5

Add relevant imports to code snippets in amp.md (#9255)

ghpvnist committed 194 days ago

Verified 364bb7eb

Revert "Switch bazel default spawn strategy to `sandboxed`" (#9212)

zhanyong-wan committed 194 days ago

Verified aa44bdea

changed build to mostly match upstream PT, fixed some doc references. (#9248)

yaoshiang committed 194 days ago

Verified 066e69ed

Fix unspecified behavior on custom calls. (#9247)

ysiraichi committed 195 days ago

Verified f0707217

Delete unused tracing functions after codegen migration. (#9240)

ysiraichi committed 196 days ago

Verified 45bff9b8

Fix docstring indentation error (#9244)

ghpvnist committed 199 days ago

Verified b6d8468c

Modify NoneRemover to return modified list/tuple (#9232)

ghpvnist committed 199 days ago

Verified 17401545

Separate pallas test into a separate test shard (#9231)

ghpvnist committed 199 days ago

Verified 0d14c8ae

Configure jobs to use better machine (#9229)

ghpvnist committed 199 days ago

Verified 2cf010bd

Fix + Run `DynamicShapeDetector` tests on CI. (#9075)

ysiraichi committed 199 days ago

Verified 6665085b

Older