pytorch/xla

Pull Requests Commits

enable_python_dispatcher in some XLA custom passes. (#9312)

laithsakka committed 207 days ago

Verified 9a517a95

Test torchprime from PyTorch/XLA (#9152)

tengyifei committed 207 days ago

Verified 728043e6

Extend build_developer.sh and git_sync_main.py with a -a flag. (#9319)

zhanyong-wan committed 207 days ago

Verified 37ed591d

Fix torchgen imports in codegen (#9310)

XuehaiPan committed 207 days ago

Verified d82e15c4

Stop tracking data and log files (#9308)

zzzwen committed 207 days ago

Verified 7875be0f

Support trace_me and xp.Trace in assume_pure (#9311)

tengyifei committed 207 days ago

Verified d663c7ed

Update dependency Updater (#9302)

tengyifei committed 210 days ago

Verified 6e8e7dbe

Merge cpp_tests1 and cpp_tests2 into single shard (#9298)

ghpvnist committed 210 days ago

Verified 6ed3ca15

Add a script to automatically format C++/Python files upon `git push` (#9293)

zhanyong-wan committed 210 days ago

Verified 9166877e

[torchax] Fixes test for kthvalue pytorch/xla#7458 (#9223)

qianminj123 committed 210 days ago

Verified 002f27a8

Sync up jax date with libtpu date (#9296)

tengyifei committed 211 days ago

Verified a687881a

[dynamic shapes] support sym_min, sym_max for XLA SymNode (#9291)

pianpwk committed 211 days ago

Verified e408a33c

Misc torchax improvements (#9294)

qihqi committed 211 days ago

Verified 6bab9e2c

Do not clear UncachedCompile between graph executions (#9282)

haifeng-jin committed 211 days ago

Verified c7a96713

Pin update 06052025 (#9299)

bhavya01 committed 211 days ago

Verified 01072eb2

Deprecate XLA:CUDA: add warning on initialization. (#9295)

ysiraichi committed 211 days ago

Verified 70de5681

Split CPU tests so the shards have a more even test time (#9292)

ghpvnist committed 211 days ago

Verified 01decbbd

Add cache to `value_and_grad_partitioned` (#9163)

iwknow committed 211 days ago

Verified d1ddeed6

Introduce annotate_custom_sharding binding (#9203)

rpsilva-aws committed 211 days ago

Verified 402612d1

Deprecate `devkind` field from `xla_model.xla_device` (#9284)

ghpvnist committed 212 days ago

Verified 5771cd31

Add w8a8 quantized matmul torchxla wrapper (#9290)

vanbasten23 committed 212 days ago

Verified a6f2b27d

Add a script to update local/forked repos to match upstream (#9288)

zhanyong-wan committed 212 days ago

Verified bc8194e0

Add w8a8 quantized matmul kernel (#9278)

vanbasten23 committed 212 days ago

Verified 25048887

[Kernel] add heuristic gmm block sizes choosing logic (#9289)

yaochengji committed 212 days ago

Verified e51af253

Disable AMP by default on CPU (#9218)

haifeng-jin committed 212 days ago

Verified 275f6e92

Split training tests to separate test shard (#9281)

ghpvnist committed 213 days ago

Verified 3e556dc8

Remove a few CUDA tests in preparation for #9202 (#9286)

ghpvnist committed 213 days ago

Verified 97b75eab

Support fp8 lowering in scalar (#9283)

yaochengji committed 213 days ago

Verified d18fc759

[torchax] Make copy_ works on different device. (#9211)

qihqi committed 214 days ago

Verified c71ff48b

Remove GPU tests from test/run_tests.sh (#9280)

zhanyong-wan committed 214 days ago

Verified b177f3d3

Newer Older