strisunshinewentingwang/pytorch

Pull Requests Commits

Remove `THTensor_(fill)` & `THTensor_(zero)` (#39727)

ShawnZhong committed 4 years ago

be838504

[Doc] Clarify that variance estimor is biaised for normalization layers (#39752)

Antoine Broyelle committed 4 years ago

bfa76ff4

Throws runtime error when performing integer division using torch.div (#38620)

Mike Ruberry committed 4 years ago

95489b59

[vulkan][CI] CI android build abi x86 with USE_VULKAN (#39767)

IvanKobzarev committed 4 years ago

cb519801

[vulkan][build_fix] Fix Vulkan Build; Prepacking uses new register api (#39771)

IvanKobzarev committed 4 years ago

a5fbd3ef

Peel Loop (#39434)

Krovatkin committed 4 years ago

7f55197a

Fix parsing of subscript expressions using python resolver (#39269)

wconstab committed 4 years ago

a1071e5d

Remove `MultiheadAttention` weights from constants (#39768)

jungomi committed 4 years ago

6748fbd3

Add rpc.async_execution support for rpc.remote on script functions (#39758)

mrshenli committed 4 years ago

3fb1e73a

[quantization] Remove duplicated piece of code in test (just a nit). (#39770)

Radhakrishnan Venkataramani committed 4 years ago

eb7843ed

Refactor CUDA bernoulli_kernel by using uniform_and_transform (#39652)

Kiyosora committed 4 years ago

d04a3fcc

Update TensorPipe submodule (#39783)

lw committed 4 years ago

68b87406

[JIT] Add Type::repr_str to return human-readable str (#39544)

Yanan Cao committed 4 years ago

c22bbb21

[Easy Review] Fix ProcessGroupRpcBackendOptions Doc (#39787)

mrshenli committed 4 years ago

4e892bd9

[quant][graphmode] Support quantization for `aten::append` (#39644)

jerryzh168 committed 4 years ago

7994d6e1

Remove unnecessary !op.is_read_write test in compute_names/compute_shape. (#39747)

ezyang committed 4 years ago

08105a00

always resize_ min/max outputs (#39696)

Natalia Gimelshein committed 4 years ago

b1750cb8

[ROCm] explicitly embed version within image name (#39735)

jeffdaily committed 4 years ago

9ba2530d

[quant][graphmode] Fix a corner case in handling `if` in insert_observers (#39615)

jerryzh168 committed 4 years ago

2d589bc9

Changes TensorIterator computation to not consider out kwarg, lets UnaryOps safe cast to out (#39655)

Mike Ruberry committed 4 years ago

0aecbbb7

[PyTorch] Make DDP reducer work under distributed autograd (#37998)

HC Zhu committed 4 years ago

acc13ac8

correct some cpp extension code usages and documents (#39766)

lixinyu committed 4 years ago

7cb4eae8

Migrate AT_DISPATCH_COMPLEX_TYPES to c10::complex (#39564)

zasdfgbnm committed 4 years ago

be13388a

Preserve user specified attributes and methods (#38830)

bzinodev committed 4 years ago

9111ae77

[TensorExpr] Fast sigmoid for LLVM (#39717)

zheng-xq committed 4 years ago

6bdfd6ae

[iOS] Add nonVarTypeModeGuard to fix the unit test (#39743)

xta0 committed 4 years ago

30792073

[JIT] consider side effects when trying moves in alias analysis (#39497)

Elias Ellison committed 4 years ago

2193fa11

move mm_cpu from BlasWrappersCPU.cpp to LinearAlgebra.cpp and delete the former file (#39700)

husthyc committed 4 years ago

3cf9b7d9

Freezing Module containing fork subgraphs (#37044)

bzinodev committed 4 years ago

1b99be90

[quant][graphmode] Test weight observer for dynamic quant (#39687)

supriyar committed 4 years ago

0fe1ec3c

Newer Older