strisunshinewentingwang/pytorch

Pull Requests Commits

[opinfo] use dtypes instead of dtypesIfCPU (#68732)

kshitij12345 committed 2 years ago

01ddd5dd

Tune test_reference_numerics_normal (#68019)

zasdfgbnm committed 2 years ago

cffad597

[JIT][Symbolic Shape Analysis] expose op shape functions (#68748)

makslevental committed 2 years ago

5fdcc20d

Revert D32599540: [pytorch][PR] implemented 'torch.distributions.constraints.symmetric' checking if the tensor is symmetric at last 2 dimension.

ngimel committed 2 years ago

f14c16e5

partial revert of D32522826 (#68889)

Krovatkin committed 2 years ago

c2e3b92d

native_functions.yaml: remove SparseXPU which is added by accident (#68791)

guoyejun committed 2 years ago

4afa5ea0

Add slow path to `getCustomClassTypeImpl` (#68717)

malfet committed 2 years ago

c5f63f85

Revert D32650384: OpInfos for torch.{flatten, column_stack}

malfet committed 2 years ago

14dc9759

Update empty and empty_like examples in docs (#68874)

RedTachyon committed 2 years ago

96929ea9

[CUDA Pinned Memory] Event recording with non-blocking copies should track the storage context, not the tensor data pointer (#68749)

Andrew Tulloch committed 2 years ago

d44e610e

implemented 'torch.distributions.constraints.symmetric' checking if the tensor is symmetric at last 2 dimension. (#68644)

nonconvexopt committed 2 years ago

bc3bdbc8

[quant][graphmode][fx] Fork subgraph_rewriter from torch.fx to quantization (#68228)

jerryzh168 committed 2 years ago

1940cc02

OpInfos for torch.{flatten, column_stack} (#67555)

anjali411 committed 2 years ago

aceb46e4

Add docs entry for `adjoint`. (#68869)

lezcano committed 2 years ago

cf544169

OpInfos for torch.atleast_{1d, 2d, 3d} (#67355)

anjali411 committed 2 years ago

c7d5e0f5

Avoid dtype mismatch error in `torch.save` if storages are unallocated (#68787)

kurtamohler committed 2 years ago

b69155f7

Revert D32633806: Sparse CSR CUDA: Add block torch.addmm when mat1 is sparse

malfet committed 2 years ago

208e109d

[nnc][quantization] quantized ops for BI bytedoc via aten (#68790)

IvanKobzarev committed 2 years ago

7802953d

fix sccache issue on Windows CPU (#68870)

mszhanyi committed 2 years ago

31d36fd3

Remove extraneous logging (#68830)

H-Huang committed 2 years ago

be7e159e

[nnc] llvm_codegen quantization types for vectype (#68736)

IvanKobzarev committed 2 years ago

7d8a79b6

Sparse CSR CUDA: Add block torch.addmm when mat1 is sparse (#68707)

IvanYashchuk committed 2 years ago

b28ddd72

Cleanup old TD logic (#68842)

malfet committed 2 years ago

b5b62b34

[SR] Use std::vector::reserve for StaticModule constants (#68834)

donaldong committed 2 years ago

d9f3feb5

Update Documentation to Make CUDA Call Explicit (#67973)

gauenk committed 2 years ago

8fb9ce49

[Quant] Refactor handling of FixedQParams operators (#68143)

andrewor committed 2 years ago

79b67d9a

All get_attr node to be in64 type (#68818)

Shirong Wu committed 2 years ago

998daf44

[BE] Simplify magma installation logic (#68778)

malfet committed 2 years ago

78dce417

Fix `test_svd_errors_and_warnings` warning message when cuda >= 11.5 (#68683)

xwang233 committed 2 years ago

2cd48d14

Revert D32611368: [pytorch][PR] Initial version of general convolution_backward

osalpekar committed 2 years ago

8e343ba5

Newer Older