strisunshinewentingwang/pytorch

Pull Requests Commits

[dtensor] remove custom dispatch op (#95629)

wanchaol committed 3 years ago

e9c49049

[DCP] Expose create_read_items_for_chunk_list helper. (#97570)

Rodrigo Kumpera committed 3 years ago

342ed037

[FSDP][7/N] Add alignment padding for `use_orig_params=True` (#97667)

Andrew Gu committed 3 years ago

1c15cd48

[FSDP][6/N] Rename param/module name helpers for clarity (#97666)

Andrew Gu committed 3 years ago

b9049a7f

[FSDP][5/N] Lift `FSDPParamInfo` to use `FlatParamHandle` (#97665)

Andrew Gu committed 3 years ago

30a6ed34

[FSDP][4/N] Document `use_orig_params: bool` (#97664)

Andrew Gu committed 3 years ago

5d554ca2

[FSDP][3/N] Minor fixes (rename, assert message) (#97663)

Andrew Gu committed 3 years ago

c6225599

[FSDP][2/N] Rename "flattened parameter" -> "flat parameter" (pt. 2) (#97662)

Andrew Gu committed 3 years ago

a27882ec

[FSDP][1/N] Rename "flattened parameter" to "flat parameter" (#97661)

Andrew Gu committed 3 years ago

bd979737

[Vulkan] Merge upsample_nearest2d and quantized_upsample_nearest2d (#97467)

manuelcandales committed 3 years ago

2bca64ae

[CI] Run benchmark test with dynamo_eager in periodic (#97543)

desertfire committed 3 years ago

a9a81ab7

remove dead torch_pb.h library (#97599)

mikey dagitses committed 3 years ago

82592f7e

Added ModuleInfos for {*}LU modules (#97375)

mikaylagawarecki committed 3 years ago

a283c15e

[eazy][inductor] fix typo in mm max-autotune log (#97486)

shunting314 committed 3 years ago

1c3ec7c4

SymIntify maybe_multiply (#97675)

ezyang committed 3 years ago

32fdd445

dont bake in defaults when tracing *_like factories (#97564)

bdhirsh committed 3 years ago

35c9ea89

make_fx, make pre_autograd a kwarg (#97559)

bdhirsh committed 3 years ago

2ca911f2

Allow -ic when no pending jobs (#97707)

Catherine Lee committed 3 years ago

6c450c78

[inductor] use torch.prifiler in the triton wrapper (#97405)

shunting314 committed 3 years ago

652592ef

Run _calculate-docker-image on 2xlarge with a larger disk space (#97551)

huydhn committed 3 years ago

6c43e9fd

[easy] Update xla hash pin merge rule (#97700)

Catherine Lee committed 3 years ago

7d944933

remove dead proto_convert library (#97598)

mikey dagitses committed 3 years ago

5d33596c

Fix typos under torch/distributed directory (#95638)

kiszk committed 3 years ago

35fd5c54

With Chillee's permission, add me to all Chillee's diffs (#97632)

ezyang committed 3 years ago

8313bece

Apparently aot_function doesn't cache anymore (#97610)

ezyang committed 3 years ago

6430dad7

Make test_binary_shape_functions actually test the ops (#90566)

kit1980 committed 3 years ago

b24052b1

[Vulkan] Fix broadcasting in quantized elementwise ops (#97554)

manuelcandales committed 3 years ago

b66a121c

Add decomposition for aten.squeeze.dims op (#97020)

vivekkhandelwal1 committed 3 years ago

5da86bbb

Sort: Use cub::WarpMergeSort for small sorts (32 < n <= 128) (#96223)

peterbell10 committed 3 years ago

91cce4c0

Add ModuleInfos for Adaptive{Max/Avg}Pool ops (#97291)

mikaylagawarecki committed 3 years ago

236bac81

Newer Older