strisunshinewentingwang/pytorch

Pull Requests Commits

support negative axis in concat and split operators

PenghuiCheng committed 5 years ago

74f6c55f

Revert "Redefine scheduler to set learning rate using recursive formula" #14010 (#21463)

ezyang committed 5 years ago

3889855a

Add a 'dim' argument to nuclear norm (#21022)

skrah committed 5 years ago

8b9b215d

Implements divmod function (#20979)

pandeykartikey committed 5 years ago

2378c120

Uninitialized Ivalue (#21387)

eellison committed 5 years ago

8a88d331

Use schema string specification in derivatives.yaml. (#20916)

gchanan committed 5 years ago

dd0ffd68

Allow tensors with requires_grad=True in c10 ops (#21599)

smessmer committed 5 years ago

5f25a252

Revert D15717575: [pytorch][PR] Fix bug in multinomial_alias_draw

colesbury committed 5 years ago

5a48642f

fix optional type promotion for classes (#21593)

suo committed 5 years ago

4fb302eb

Consider contained types in alias analysis (#21431)

suo committed 5 years ago

a436822c

cleanups to memory_dag (#21430)

suo committed 5 years ago

bb4aff26

cleanups to alias analysis interfaces (#21397)

suo committed 5 years ago

ae144032

avoid calling front() on empty working set (#21396)

suo committed 5 years ago

ddac8da8

Fix bug in multinomial_alias_draw (#21324)

vishwakftw committed 5 years ago

bb1dbdb9

Krovatkin committed 5 years ago

30d69330

Skip triangular_solve CUDA test on non-default stream

vishwakftw committed 5 years ago

3df5a46a

fix test (#21594)

Elias Ellison committed 5 years ago

6f99bcda

clip sigmoid to prevent transforms return inf/nan values (#20288)

fehiepsi committed 5 years ago

91ea2cd5

Add python binding to deserialize blob (#21532)

Haixin Liu committed 5 years ago

4bdbd30b

Change compiler to use Load/Stores, then transform to SSA (#21101)

Elias Ellison committed 5 years ago

e4fae884

update hub doc (#21568)

Ailing Zhang committed 5 years ago

1e6c99a6

Don't leak threads on exit (#21438)

mal committed 5 years ago

f308b07e

fix concat and split tensor inference function (#21382)

jspark1105 committed 5 years ago

c294d64e

Documentation for locking discipline in engine.cpp/.h (#21548)

Malvika Joshi committed 5 years ago

9deab0cf

Add named_guard to native_functions options (#21373)

zou3519 committed 5 years ago

547fcaa9

Add unique_ptr<NamedTensorMeta> field to TensorImpl (#21341)

zou3519 committed 5 years ago

8ffcbfb7

Fix NVTX path on Windows

peterjc123 committed 5 years ago

f9c4d0d7

Regularization is not supported in FP16 (#21319)

Xing Wang committed 5 years ago

c4e0d616

Enabled _th_ixor_ and _th_equal for bool (#21538)

izdeby committed 5 years ago

b1bf16ee

Update module.py

ypwhs committed 5 years ago

e447a733

Newer Older