strisunshinewentingwang/pytorch

Pull Requests Commits

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

Switch most Python RAII guard usages to context manager (#102642)

zou3519 committed 1 year ago

74f10b9e

Use safe_is_leaf to test leafness (#102706)

ezyang committed 1 year ago

dcf0c5fb

add storage dtype for custom device (#102481)

heidongxianhua committed 1 year ago

d9c8f9a0

inductor: eliminate meaningless copy (#102089)

blzheng committed 1 year ago

e59db086

[Quant][PT2E][Inductor] Lower quantized conv to Inductor (#101164)

Xia-Weiwen committed 1 year ago

ce9923a1

add foreach support for custom device (#102047)

heidongxianhua committed 1 year ago

b088ff46

[Dynamo] Correct UserDefinedObjectVariable.var_getattr on function/method type (#102580)

yanboliang committed 1 year ago

9fa82c90

Newer Older