strisunshinewentingwang/pytorch

Pull Requests Commits

Back out "[pytorch][PR] Fix many type mismatches in the CUDA version of calc_digamma and calc_trigamma" (#26444)

Mike Ruberry committed 5 years ago

495dbacf

Remove quantizeBias (#26388)

jerryzh168 committed 5 years ago

fb28014a

Exclude libfbjni.so from pytorch_android not to have its duplicating (#26382)

IvanKobzarev committed 5 years ago

6387ffab

tvm operator dynolog (#26295)

ajyu committed 5 years ago

e44ea6cd

Move the CUDA implementation of rsqrt to ATen. (#25285)

VitalyFedyunin committed 5 years ago

36ade9aa

C++ API parity: at::Tensor::is_leaf

pbelevich committed 5 years ago

44ffbc43

fix composite learning rate (#26227)

olittle committed 5 years ago

a8386d2a

Add extra filtering for scale/zero_point/dtype in FoldQuantizeCallIntoBuffer (#26224)

jerryzh168 committed 5 years ago

f75c1e49

Adding quantized::conv2d function for pytorch mobile in c10 (#26152)

supriyar committed 5 years ago

b23be955

remove extra get_worker_id call in distributed rpc init (#26381)

rohan-varma committed 5 years ago

1f510512

Add filter function to subgraph rewriter runGraph (#26223)

jerryzh168 committed 5 years ago

f29e0d70

Use static type information to restore type tags (#25447)

zdevito committed 5 years ago

12762cd5

Add ivalue::type(), part 1 (#25439)

zdevito committed 5 years ago

ad0af112

add pass for onnx scalar type conversion (#24378)

BowenBao committed 5 years ago

d02369da

Adds dtypes decorators to and allows helper methods in device generic test classes (#26375)

Mike Ruberry committed 5 years ago

248d5857

Disable QNNPACK tests if pytorch is not built with it. (#26427)

supriyar committed 5 years ago

52d999e1

Puts ROCm tests on default stream (#26394)

Mike Ruberry committed 5 years ago

a5616602

Fix many type mismatches in the CUDA version of calc_digamma and calc_trigamma (#25791)

xuhdev committed 5 years ago

13b54460

Add support for lists for prim::min and prim::max

Krovatkin committed 5 years ago

18eb92e2

Change "named_guard" in native_functions to "supports_named_tensor" (#26352)

zou3519 committed 5 years ago

76fb909b

clean up the PR job script for iOS build (#26353)

xta0 committed 5 years ago

ecb82ed5

Add a float version of calc_erfinv (by templating) on CPU (#26070)

xuhdev committed 5 years ago

2801df5b

Make ProcessGroupAgent take num_send_recv_threads as constructor argument (#26313)

xush6528 committed 5 years ago

b0b0f2c6

Removes torchtest, expands generic device testing (#26374)

mruberry committed 5 years ago

388cfdf2

use allgatherv for sparse all reduce (#23917)

zhaojuanmao committed 5 years ago

ed097048

C++ API parity: at::Tensor::grad

pbelevich committed 5 years ago

98ccae09

Fix no tab check (#26399)

houseroad committed 5 years ago

72aeafd3

Resolve #25605 cyclic reference in _LRScheduler (#25776)

huzecong committed 5 years ago

b8ae4d0f

Change '*' to '...' and `...` for named tensor API functions. (#26350)

zou3519 committed 5 years ago

bae75284

Rename torch.namedtensor -> torch._namedtensor_internals (#26349)

zou3519 committed 5 years ago

277d442d

Newer Older