pytorch/xla

Pull Requests Commits

yaochengji committed 251 days ago

119a50b6

set v4 of vmem_limit_bytes to 16MB

yaochengji committed 251 days ago

8286d566

yaochengji committed 251 days ago

56147e0f

yaochengji committed 251 days ago

faa2d2a1

yaochengji committed 251 days ago

3e54ecbe

yaochengji committed 251 days ago

e4a23760

yaochengji committed 252 days ago

437fa7d9

Add heuristic default block sizes for different cases in ragged attention kernel

yaochengji committed 252 days ago

64ac8b68

[torchax] Support linalg.det and linalg.lu_solve (#8872)

dvhg committed 252 days ago

Verified 2d6f57a3

Fix cuda dependency not found (#8903)

zpcore committed 252 days ago

Verified dbb03aa1

Lower `isneginf()`. (#8912)

ysiraichi committed 252 days ago

Verified cba9ff9f

Fix fp8 test (#8909)

lsy323 committed 253 days ago

Verified 985fa188

Add Pallas tutorials to the Pallas guide at torch_xla. (#8904)

vanbasten23 committed 253 days ago

Verified 450b4dce

point build to rc3 (#8907)

zpcore committed 253 days ago

Verified 2999e72b

Use shard_as in scan to ensure that inputs and their gradients have the same sharding (#8879)

tengyifei committed 256 days ago

Verified 6d88c089

[Build] Pin cmake to build PyTorch (#8902)

lsy323 committed 256 days ago

Verified fe3bb7f6

Remove dynamic grid (#8896)

bythew3i committed 257 days ago

Verified b24e6e95

Update cudnn dependency for newer cuda versions (#8894)

zpcore committed 257 days ago

Verified 92873377

8740 add single processing to getting started instructions (#8875)

pgmoka committed 257 days ago

Verified e341ff09

[ragged-paged-attn] Combine k_pages and v_pages on num_kv_head (#8892)

bythew3i committed 258 days ago

Verified 7a3c0518

Encapsulate Mesh invariants (#8882)

rpsilva-aws committed 258 days ago

Verified 96ad8f5f

Avoid unnecessary copy in TensorSource (#8849)

lsy323 committed 258 days ago

Verified 8dc5b496

update Dockerfile config for CI (#8886)

zpcore committed 258 days ago

Verified 3066b034

Increase CPU node count in tpu-ci (#8885)

tengyifei committed 258 days ago

Verified 6a7ce475

Cache HLO in xb.call_jax and support non-tensor args (#8878)

tengyifei committed 260 days ago

Verified a3ef52e5

Add lowering for `bitwise_right_shift`. (#8866)

ysiraichi committed 260 days ago

Verified 2aa11ccc

Correct type conversions in OpBuilder and add corresponding test. (#8873)

iwknow committed 260 days ago

Verified f2bdecfa

8727 create a site map or centralize links in readme (#8843)

pgmoka committed 260 days ago

Verified 3808df01

Add lowering for `bitwise_left_shift`. (#8865)

ysiraichi committed 260 days ago

Verified 46a7990f

Update README.md with torch_xla-2.8.0 wheels (#8870)

bhavya01 committed 263 days ago

Verified 8030f059

Older