strisunshinewentingwang/pytorch

Pull Requests Commits

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

Remove thread_local RecordFunctionGuard from profiler. (#44646)

rohan-varma committed 3 years ago

1bd6533d

CUDA BFloat16 pooling (#45151)

zasdfgbnm committed 3 years ago

67a19fec

[jit] gtestify test_argument_spec.cpp (#45019)

suo committed 3 years ago

666223df

[quant][graphmode][jit][api] Expose preserved_attrs from finalize to convert_jit (#44490)

jerryzh168 committed 3 years ago

f575df20

[JIT] Add default arguments for class types (#45098)

Meghan Lele committed 3 years ago

e0451199

[tensorexpr] Add flag to fuse with unknown shapes (#44401)

bwasti committed 3 years ago

ebde5a80

[caffe2] Support data types in shape hints (#45110)

Hao Lu committed 3 years ago

c0267c68

install ATen/native/cuda and hip headers (#45097)

jeffdaily committed 3 years ago

b98ac208

Relax CUDA architecture check (#45130)

malfet committed 3 years ago

2a37f3fd

[quant][graphmode][fx] Custom module support (#44766)

jerryzh168 committed 3 years ago

ccfbfe5e

[resubmit][FX] s/get_param/get_attr/ (#45147)

James Reed committed 3 years ago

7f4a27be

[PyTorch] Enable type check for autocast_test_lists (#45107)

hangjunxu committed 3 years ago

35cdb013

[JIT] Fix WithTest.test_with_exceptions (#45106)

Meghan Lele committed 3 years ago

cddcfde8

Newer Older