strisunshinewentingwang/pytorch

Pull Requests Commits

Revert D15523444: [jit] move casting ops from prim to aten

wanchaol committed 5 years ago

2cd1c786

Enabled min, max, minall, maxall, cmin, cmax, cmaxValue, cminValue for bool tensors (#21031)

izdeby committed 5 years ago

7cb1aa67

Assert against using Operator methods not supported when exporting it to c10, part 2 (#17946)

smessmer committed 5 years ago

85777b92

move casting ops from prim to aten (#21002)

wanchaol committed 5 years ago

a0111aaf

Add start and step parameters for range in torchscript (#20795)

Chillee committed 5 years ago

dd903eb6

Revert D15502768: [pytorch][PR] [jit] Make ScriptModule.training an attribute instead of a parameter

David Riazati committed 5 years ago

fa8c132e

fix `dequantize_linear` (#21035)

jerryzh168 committed 5 years ago

94b97060

print a warning if a type annotation prefix is invalid according to mypy (#20884)

Krovatkin committed 5 years ago

cbf2a4f5

Removed accidental TensorFlow dependency (#21066)

Dawars committed 5 years ago

a6bb1549

Hook to store additional metadata about environment (#20863)

Dmytro Dzhulgakov committed 5 years ago

f2199a34

Added possibility to index scalars by bool masks (#21030)

izdeby committed 5 years ago

00c15849

Improve test_proper_exit error printing (#20166)

ssnl committed 5 years ago

1d4685c2

ctc_loss: fix backward when 2d target tensor is larger than max_target_length (#20971)

t-vi committed 5 years ago

aa42742d

DilatedMaxPool2d: small cleanup

skrah committed 5 years ago

55f5eb3c

Port dilated_max_pool3d() to ATen

skrah committed 5 years ago

f8565121

Revert D15499749: [pytorch][PR] Add `Tensor.T` attribute to reverse dimensions

ezyang committed 5 years ago

0544a491

Remove THSTensor and SparseTensorRef (#20877)

Roy Li committed 5 years ago

3038cf8e

Fix __irshift__ dispatch (#21047)

Roy Li committed 5 years ago

9faa409b

Remove extraneous TensorId checks in as_strided (#21045)

Roy Li committed 5 years ago

8dda19b7

Fix tracing bugs where using `1 - x` in C++ would cause the size of 1 to get hardcoded. (#20932)

zdevito committed 5 years ago

d76546a4

Make build with makefiles less noisy (#21053)

bddppq committed 5 years ago

5c53aa48

Fix get_gpu_memory_info in non-cuda builds (#21054)

bddppq committed 5 years ago

9b147961

Deprecate variadic inputs of checkpoint_sequential (#21006)

sublee committed 5 years ago

ffdce790

Allow nondet_tol for nondeterminism in gradcheck and gradgradcheck (#20980)

t-vi committed 5 years ago

d23d04f1

Fix typo in CyclicLR docs (#21021)

njdalton committed 5 years ago

d190450a

add simple memory analyzer and log warning if GPU underutilized (#21024)

Aapo Kyrola committed 5 years ago

f1fe4b11

Fix warning in register_c10_ops by making index unsigned (#20964)

Chillee committed 5 years ago

1bed5f39

Enable batched QR decomposition and add a `some` option (#20689)

vishwakftw committed 5 years ago

f6ec4648

Use constants from math.h for gelu op (#20974)

xiaomengy committed 5 years ago

c1048182

tracing for intra_op_parallel (#20603)

jspark1105 committed 5 years ago

0290897b

Newer Older