strisunshinewentingwang/pytorch

Pull Requests Commits

Resubmit more code fakefp16 mapping unification (#32798)

Yinghai Lu committed 4 years ago

94ddc2c4

Centralize addition of "always on" dispatch keys. (#32734)

ezyang committed 4 years ago

690d41f2

Make DispatchKeyGuards accept DispatchKey::Undefined (#32729)

ezyang committed 4 years ago

5ddd2cd9

Rename DispatchKey::UndefinedTensorId to Undefined (#32728)

ezyang committed 4 years ago

3d0a470d

Remove GIL from RRefContext (#32807)

mrshenli committed 4 years ago

a40a19cc

Fixes moving after weight norm application (#32563)

mruberry committed 4 years ago

413c0f6c

Make python version a parameterizable option for Windows CI.

peterjc123 committed 4 years ago

9bab617b

Fix backcompat for linear_relu_dynamic_fp16 (#32803)

James Reed committed 4 years ago

cc35c876

Re-enable non-deterministic autograd tests

albanD committed 4 years ago

fa658592

Removing @expectedFailureXLA from test_nll_loss_empty_tensor_reduction_mean (#32701)

pbelevich committed 4 years ago

85bd3e5b

Revert D19611800: [PyTorch][TorchScript] Add support for join on List of strings in TorchScript

ezyang committed 4 years ago

68742789

Reduce RPC branches for Python/BuiltinOp/TorchScript (#32689)

xush6528 committed 4 years ago

b0923acb

Fix/simplify alias annotation handling in op codegen. (#32574)

Basil Hosmer committed 4 years ago

affd598c

Some work on eager op binding codegen (gen_python_functions.py) (#29986)

Basil Hosmer committed 4 years ago

fb159b52

[pytorch] Minor: avoid acquiring GIL twice in PyRRef::localValue() (#32785)

jjlilley committed 4 years ago

821b6aa7

Add support for Dynamic LSTM quantization on Mobile (#32757)

supriyar committed 4 years ago

c2d736ce

Fixed access to element in size tensor for scripting (#32652)

Brian Stark committed 4 years ago

55c382e6

[PyTorch][TorchScript] Add support for join on List of strings in TorchScript

mvsampath committed 4 years ago

8ead65a9

Resolve rendezvous race condition

mingzhe09088 committed 4 years ago

cccf5e70

[jit] fix the NoneType param/buffer hack (#32745)

suo committed 4 years ago

3552be10

enable empty batch for all flavor of convolutions (#32709)

ngimel committed 4 years ago

2e359ef8

[pytorch][embeddingbag_8bit] Add include_last_offset option to Fused 8bit EmbeddingBag and parallelize the op (#32683)

jianyuh committed 4 years ago

a840afbe

Logspace fixes (#32744)

Enealor committed 4 years ago

b565d9b3

[quantization] Remove incorrect fp16 dynamic linear/relu op

James Reed committed 4 years ago

fc2ff791

Remove -Werror from test/cpp_extensions/setup.py (#32704)

pbelevich committed 4 years ago

9357b911

Fix ivalue_inl.h:353:29: warning: comparison of unsigned expression >= 0 is always true (#32778)

zasdfgbnm committed 4 years ago

8b187e8f

Revert D19597036: More code fakefp16 mapping unification

ezyang committed 4 years ago

c47c78d0

Refreshing numel on a stride update is pointless. (#32116)

ezyang committed 4 years ago

3ee6673e

Delete resize_dim() (#32114)

ezyang committed 4 years ago

8c6f52ac

Expunge last two sites of resize_dim (#32112)

ezyang committed 4 years ago

b371eab8

Newer Older