strisunshinewentingwang/pytorch

Pull Requests Commits

Create experimental FX graph manipulation library (#44775)

gcatron committed 3 years ago

ef414725

Quantized operators template selective (#45509)

iseeyuan committed 3 years ago

d6429928

fix standard deviation gradient NaN behavior (#45468)

Randall Hunt committed 3 years ago

ab5cf16b

Update backward formula for torch.dot and add backward definition for torch.vdot (#45074)

anjali411 committed 3 years ago

18876b57

Add docs to a pytorch.github.io/doc/tag directory when repo is tagged (#45204)

mattip committed 3 years ago

147c88ef

Updates nonzero's as_tuple behavior to no longer warn. (#45413)

Mike Ruberry committed 3 years ago

b66ac1e9

Remove unnecessary __at_align32__ in int_elementwise_binary_256 (#45470)

xuhdev committed 3 years ago

0df99ad4

Move mobile specific CPUCachingAllocator to c10/mobile folder. (#45364)

kimishpatel committed 3 years ago

6e55a26e

Updates deterministic flag to throw a warning, makes docs consistent (#45410)

Mike Ruberry committed 3 years ago

b2925671

Conservative-ish persistent RNN heuristics for compute capability 8.0+ (#43165)

mcarilli committed 3 years ago

aa2bd7e1

Updated `cholesky_backward` for complex inputs (#45267)

IvanYashchuk committed 3 years ago

f47fd0eb

Support bfloat16 and complex dtypes for logical_not (#43537)

xuhdev committed 3 years ago

15f85eea

Fix model_name not logged properly issue. (#45488)

xcheng16 committed 3 years ago

ea59251f

[JIT] Enable @unused syntax for ignoring properties (#45261)

Meghan Lele committed 3 years ago

09b3e16b

Add mobile_optimized tag to optimized model. (#45479)

Akshit Khurana committed 3 years ago

5f49d14b

[vulkan][android][test_app] Add test_app variant that runs module on Vulkan (#44897)

IvanKobzarev committed 3 years ago

17be7c6e

[android][vulkan] Module load argument to specify device cpu/vulkan (#44896)

IvanKobzarev committed 3 years ago

2c300fd7

Reorganized Sorting.cpp method order (#45083)

heitorschueroff committed 3 years ago

fe9019cb

Revert D23789657: [wip] fast typeMeta/ScalarType conversion approach 2

Mike Ruberry committed 3 years ago

ab5edf21

Enable torch.cuda.amp typechecking (#45480)

malfet committed 3 years ago

b3135c20

Add cusolver guard for cuda >= 10.1.243 (#45452)

xwang233 committed 3 years ago

df0de780

Improves fft doc consistency and makes deprecation warnings more prominent (#45409)

Mike Ruberry committed 3 years ago

bb19a554

Auto set libuv_ROOT env var for Gloo submodule on Windows platform (#45484)

gunandrose4u committed 3 years ago

0a38aed0

Makes rdiv consistent with div (#45407)

Mike Ruberry committed 3 years ago

6d37126a

Add check for Complex Type to allow non integral alpha. (#45200)

hlahkar committed 3 years ago

7cde662f

Optimize view_as_complex and view_as_real (#44908)

peterbell10 committed 3 years ago

0806c58e

Updates torch.floor_divide documentation to clarify it's actually torch.trunc_divide (or torch.rtz_divide) (#45411)

Mike Ruberry committed 3 years ago

87f98a5b

Missing tests about torch.xxx(out=...) (#44465)

antocuni committed 3 years ago

37f9af7f

Revert D23966878: [pytorch][PR] This PR flips a switch to enable PE + TE

Mike Ruberry committed 3 years ago

56af1226

[wip] fast typeMeta/ScalarType conversion approach 2 (#44965)

Basil Hosmer committed 3 years ago

1ed1a2f5

Newer Older