strisunshinewentingwang/pytorch

Pull Requests Commits

[quant][pyper] make embedding_bag quantization static (#44008)

supriyar committed 4 years ago

164b96c3

[quant] Support aten::embedding_bag quantization in graph mode (#43989)

supriyar committed 4 years ago

a0ae416d

Add const to getTensors method of GradBucket. (#44126)

Yi Wang committed 4 years ago

15a73681

[JIT] Remove references to no longer generated _tanh_backward and _sigmoid_backward (#44138)

Elias Ellison committed 4 years ago

5bd29027

[TensorExpr fuser] Guard nodes that have tensor output properties determined by non-tensor inputs (#44137)

Elias Ellison committed 4 years ago

df67f0be

Further expand coverage of addmm/addmv, fix 0 stride (#43980)

zasdfgbnm committed 4 years ago

5a0d65b0

Revert D23490149: [pytorch][PR] Compile less legacy code when BUILD_CAFFE2 is set to False

wanchaol committed 4 years ago

d07a36e0

fx quant prepare: clarify naming (#44125)

vkuzo committed 4 years ago

618b4dd7

torchscript graph mode quant: remove benchmark filter (#44165)

vkuzo committed 4 years ago

a940f5ea

[dper3] replace LengthsGather lowlevel module's PT implemetnatio to use caffe2 op

ashishenoyp committed 4 years ago

8c64bb4f

[PyTorch][Mobile] Insert the module name as `name()` to metadata dict if metadata doesn't contain "model_name" (#44227)

husthyc committed 4 years ago

398409f0

Compile less legacy code when BUILD_CAFFE2 is set to False (#44079)

malfet committed 4 years ago

15e99b6f

[ONNX] Update repeat op (#43430)

shubhambhokare1 committed 4 years ago

f3bf6a41

[DPER3] AOT integration

ChunliF committed 4 years ago

3699274c

Add remote_parameters() into RemoteModule class. (#43906)

Yi Wang committed 4 years ago

8b17fd25

[BUILD] Guard '#pragma unroll' with COMPILING_FOR_MIN_SIZE

alx32 committed 4 years ago

8f37ad82

[ONNX] Enable new scripting passes for functionalization and remove_mutation (#43791)

neginraoof committed 4 years ago

3d7c22a2

[FX] Fix forward merge conflict breakage (#44221)

James Reed committed 4 years ago

70bbd084

Fix potential divide by zero for CostInferenceForRowWiseSparseAdagrad

glennq committed 4 years ago

4562b212

[fx] get rid of graph_module.root (#44092)

zdevito committed 4 years ago

2ad5a82c

Revert D23468286: [pytorch][PR] Optimize code path for adaptive_avg_pool2d when output size is (1, 1)

ngimel committed 4 years ago

0c2bc4fe

Revert D23503636: [pytorch][PR] [NNC] make inlining immediate (take 2) and fix bugs

Mikhail Zolotukhin committed 4 years ago

6474057c

[ONNX] Fix split export using slice (#43670)

neginraoof committed 4 years ago

539d029d

[FX] __str__ for GraphModule and Graph (#44166)

James Reed committed 4 years ago

af13faf1

[pytorch] remove code analyzer build folder between builds (#44148)

ann-ss committed 4 years ago

0e3cf6b8

Updates to SCCACHE for ROCm case (#44155)

pruthvistony committed 4 years ago

f38e7aee

replace the white list from default mappings (#41802)

VinodSKumar committed 4 years ago

2a1fc566

Control NCCL build parallelism via MAX_JOBS environment var (#44167)

malfet committed 4 years ago

4d431881

Limit MAX_JOBS to 18 for linux binary builds (#44168)

malfet committed 4 years ago

6aba58cf

Enable test_cublas_config_deterministic_error for windows (#42796)

guyang3532 committed 4 years ago

6cecf7ec

Newer Older