strisunshinewentingwang/pytorch

Pull Requests Commits

[TensorExpr] When lanes differ, insert Broadcast instead of Cast (#45179)

asuhan committed 3 years ago

76c185dc

[quant][eagermode] Custom module support (#44835)

jerryzh168 committed 3 years ago

f93ead6d

[TensorExpr] Disallow arithmetic binary operations on Bool (#44677)

asuhan committed 3 years ago

04959988

[TensorExpr] Remove unused EvalConstExpr function (#45180)

asuhan committed 3 years ago

8e0fc711

Fix typos. (#45195)

Yi Wang committed 3 years ago

2a1a51fa

gtestify dce and fuser tests (#45055)

suo committed 3 years ago

246bd942

gtest-ify JIT tests, through the letter c (#45020)

suo committed 3 years ago

d2b04503

[jit] allow submodule methods inference rule be different (#43872)

wanchaol committed 3 years ago

3f89b779

[NNC] Fix a bug in SplitWithMask when splitting multiple times (#45141)

nickgg committed 3 years ago

9e206ee9

[quant][graphmode][fx] qconfig_dict support more types of configurations (#44856)

jerryzh168 committed 3 years ago

adb2b380

Remove expensive call to PyObject_GetAttrString in PyTorch_LookupSpecial (#44684)

bradleyhd committed 3 years ago

21fabae4

Dockerfile: Support CUDA 11 (#45071)

tindzk committed 3 years ago

99242eca

Fix inlining interface call in fork subgraph (#43790)

bzinodev committed 3 years ago

4d80c8c6

Make cudaHostRegister actually useful on cudart. (#45159)

ezyang committed 3 years ago

da4033d3

Warn if `import torch` is called from the source root. (#39995)

Taylor Robie committed 3 years ago

a5a4924c

Update true_divide_out to use at::. (#45079)

Ailing Zhang committed 3 years ago

9db38712

Filter `strtod_l` is undeclared errors from sccache log (#45183)

malfet committed 3 years ago

9e30a766

Added support for complex input for Cholesky decomposition (#44895)

IvanYashchuk committed 3 years ago

5b20bf4f

Let rpc._all_gather use default RPC timeout (#44983)

mrshenli committed 3 years ago

94c3cdd9

[PyTorch Mobile] Move string op registrations to prim and make them selective (#44960)

iseeyuan committed 3 years ago

e5bade7b

[RPC] Infer backend type if only options are given (#45065)

lw committed 3 years ago

76dc50e9

[TensorExpr] Fix operator order in combineMultilane (#45157)

asuhan committed 3 years ago

21567957

[quant][fx][bug] Fix error in convert step for QAT (#45050)

supriyar committed 3 years ago

7fba30c2

CUDA BFloat16 batched gemm (#45167)

zasdfgbnm committed 3 years ago

144dacd8

[JIT] Do not allow creating generics with None types (#44958)

malfet committed 3 years ago

989d877c

[reland][pytorch] refine dispatch keys in native_functions.yaml (1/N) (#45137)

ljk53 committed 3 years ago

0a9ac98b

[packaging] rstrip fix (#45166)

zdevito committed 3 years ago

25ed739a

torch.package - a way to package models and code (#45015)

zdevito committed 3 years ago

cb75adde

[RPC profiling] Don't wrap toHere() calls with profiling (#44655)

rohan-varma committed 3 years ago

d4a634c2

[RPC profiling] Allow disableProfiler() to be called from another thread. (#44653)

rohan-varma committed 3 years ago

70d2e4d1

Newer Older