strisunshinewentingwang/pytorch

Pull Requests Commits

Inductor cpp wrapper: cache the wrapper (#89743)

chunyuan-w committed 1 year ago

4c9992d5

[CI] Move libtorch-debug CUDA build to CUDA-12.1 (#102756)

malfet committed 1 year ago

0b732031

Fix aten.copy device mismatch bug in FakeTensor (#102664)

williamwen42 committed 1 year ago

da963d79

Add print statements to debug sharding error (#102713)

clee2000 committed 1 year ago

c7873522

Allow ORT backend for DTensor (#101914)

ashari4 committed 1 year ago

cf0aa380

Revert "[Pytorch] Add Vulkan support for aten::unsqueeze, 1d->2d, 3d->4d (#102042)"

pytorchmergebot committed 1 year ago

72ed22e8

Revert "[quant][test] Fix broken PT2 import, add warnings (#102644)"

pytorchmergebot committed 1 year ago

8b03a59e

initialize max_stream_priorities in getStreamFromPool(bool) (#102739)

842974287 committed 1 year ago

f15af198

Add `-debug` suffix to trunk libtorch builds (#102764)

malfet committed 1 year ago

67792e17

Use int64_t for indexing in `multi_tensor_apply` (#101760)

crcrpar committed 1 year ago

401109a2

check users are actually recieved in upload to s3 (#102760)

PaliC committed 1 year ago

b8e2e0e9

Skip test test_triton_bsr_dense_bmm if not TEST_WITH_TORCHINDUCTOR [v2] (#102660)

xwang233 committed 1 year ago

6340aa5d

[BE] Make `test_no_triton_on_import` simple (#102674)

malfet committed 1 year ago

ca470fc5

Fix string concatenation with non-string (#102728)

ezyang committed 1 year ago

90b1b17c

[C10D] Implement Store fallbacks for append, multi_get and multi_set. (#100768)

Rodrigo Kumpera committed 1 year ago

ca1c1fdc

[inductor] Fix a cpp wrapper codegen issue for _scaled_dot_product_efficient_attention (#102624)

desertfire committed 1 year ago

59532bd6

Serialize pytree to string v2 (#102708)

angelayi committed 1 year ago

bd0a4e2d

Revert "[Executorch][codegen] Add ETKernelIndex for aggregating all kernels for kernel (#102565)"

malfet committed 1 year ago

fb072905

[split cat fx passes] Normalize squeeze (#102294)

devashishshankar committed 1 year ago

9d9ce19d

[quant][test] Fix broken PT2 import, add warnings (#102644)

andrewor14 committed 1 year ago

f18b9f86

Remove deprecated HIP flags (#102271)

jataylo committed 1 year ago

87c976b6

[functorch] Get test_functionalize to run on FB infra (#102695)

bertmaher committed 1 year ago

30558c28

Mark job as unstable dynamically (#102426)

huydhn committed 1 year ago

08150ee0

[quant][be] Change return type for zero_point to be int32 Tensor (#102234)

jerryzh168 committed 1 year ago

ce8d3155

[Pytorch] Add Vulkan support for aten::unsqueeze, 1d->2d, 3d->4d (#102042)

lucylq committed 1 year ago

c9ae705a

[inductor] Reduce duplication of reduction combine functions (#99661)

peterbell10 committed 1 year ago

2f96981e

[quant][pt2e][be] Add QuantizationSpecBase (#102582)

jerryzh168 committed 1 year ago

d930bfc4

Back out "Add PyObject preservation for UntypedStorage (#97470)" (#102553)

842974287 committed 1 year ago

68550535

[C10D] Rewrite TCPStore client send path to minimize amount of syscalls. (#100742)

Rodrigo Kumpera committed 1 year ago

32360b48

Revert "add foreach support for custom device (#102047)"

pytorchmergebot committed 1 year ago

9d77949b

Newer Older