strisunshinewentingwang/pytorch

Pull Requests Commits

[pytorch] add flag for autograd ops to mobile builds (#43154)

ann-ss committed 4 years ago

0dc41ff4

[quant] Add benchmakrs for embedding_bag coversion ops (#43291)

supriyar committed 4 years ago

4fc9e958

streamline stride propagation logic in TensorIterator (#42922)

ngimel committed 4 years ago

c8bc298d

.circleci: Remove manual docker installation (#43277)

seemethere committed 4 years ago

ca9d4401

.circleci: Don't quote glob for conda upload (#43297)

seemethere committed 4 years ago

66a79bf1

Make _compute_linear_combination.out a true out function (#43272)

ezyang committed 4 years ago

397325a1

Increase deadline time for load_save tests (#43205)

Sean Lynch committed 4 years ago

f9a766bb

Nightly Pull (#43294)

scopatz committed 4 years ago

a2ae2d32

Fix ASAN error in QNNPACK's integration of qlinear_dynamic. (#41967)

Ashkan Aliabadi committed 4 years ago

6a09df99

Update torch.Tensor.is_set_to documentation (#43052)

subramen committed 4 years ago

60b524f2

Make Windows CUDA-11 tests master only (#43234)

malfet committed 4 years ago

4e964f3b

refactor torch/cuda/nccl.h to remove direct dependency on NCCL in libtorch_python (#42687)

Rong Rong committed 4 years ago

3eb31325

[NCCL] Changed FutureNCCL's then callback logic for better efficiency. (#42869)

sinannasir committed 4 years ago

6e1127ea

Modify Circle CI script to upload test report for analysis. (#43180)

Rong Rong committed 4 years ago

97d62bcd

[vulkan] fix invalid memory op and tests (#43312)

IvanKobzarev committed 4 years ago

0617156f

[quant][cleanup]test_qlinear_legacy should be under TestDynamicQuantizedLinear. (#40084)

Radhakrishnan Venkataramani committed 4 years ago

aad1ff9f

[jit] fix str -> Device implicit conversions (#43213)

zdevito committed 4 years ago

410d5b95

Automated submodule update: FBGEMM (#43251)

facebook-github-bot committed 4 years ago

018b4d7a

.circleci: Simplify binary upload process (#43159)

seemethere committed 4 years ago

eb7fc2e9

[GLOO] handle empty split size (#43256)

jiayisuse committed 4 years ago

d467ac8f

Implement Tensor.to batching rule (#43206)

zou3519 committed 4 years ago

7d102980

[CircleCI] Use `canary` images until VC++ 14.27 issue is resolved (#43220)

malfet committed 4 years ago

1e248cab

Add dataclasses to base Docker images. (#43217)

ezyang committed 4 years ago

bc0e1e8e

default ice-ref to c-step (#4812)

Hector Yuen committed 4 years ago

06d43dc6

2 Bit Embedding Conversion Operator support. (#43077)

Radhakrishnan Venkataramani committed 4 years ago

fa6b34b5

Fix some mistakes in native_functions.yaml (#43156)

ezyang committed 4 years ago

ab366d0f

remove thunk fix now that ROCm CI images are >= ROCm 3.5 (#43226)

jeffdaily committed 4 years ago

27ec91b0

update path in CI script to access ninja (#43236)

jeffdaily committed 4 years ago

8094228f

Optimize linux CI build/test matrix (#43240)

malfet committed 4 years ago

7c923a10

In copy_weights_to_flat_buf_views() explicitly construct tuple (#43244)

malfet committed 4 years ago

e41ca2d9

Newer Older