strisunshinewentingwang/pytorch

Pull Requests Commits

Revert "Mark protobuf include path as system include (#23012)"

soumith committed 4 years ago

23df957e

Clean up JavaDoc comments in pytorch_android

dreiss committed 4 years ago

a7b161c0

Various cleanups to pytorch_android API (#27454)

dreiss committed 4 years ago

6bae48c1

Refactor python_android test to separate Android-specific components (#27453)

dreiss committed 4 years ago

c2489437

Modify PyTorch's integration of NNPACK to use a unified underlying thread pool implementation. (#27547)

ljk53 committed 4 years ago

e058a37f

Add missing Optional annotation. (#27557)

ailzhang committed 4 years ago

aa7112a6

#include <stdexcept> into flat_hash_map.h (#27480)

pbelevich committed 4 years ago

b728ffab

update (#27386)

driazati committed 4 years ago

d67898a9

Revert to align_corners=True as default. (#27469)

ailzhang committed 4 years ago

9a256734

Fix native ctc_loss gradient indexing bug for large target sizes

t-vi committed 4 years ago

17613ad7

[android][torchvision] Add methods to write image tensor content to buffer (#27407)

IvanKobzarev committed 4 years ago

beaae6a2

MovingAverage Observer (#27396)

raghuramank100 committed 4 years ago

328f4996

Replacing the skip_list with white_list in the qconfig propagation

z-a-f committed 4 years ago

7f9096f8

Avoid calling tensor.numel() in for loops (#27298)

supriyar committed 4 years ago

7e94ee23

Factored out the default mappings

z-a-f committed 4 years ago

318fb8e8

Fix reprs for _intrinsic modules

James Reed committed 4 years ago

43bb1b23

Allow set for qconfig for dynamic_quantize

James Reed committed 4 years ago

87fbd27c

Rename _intrinsic to intrinsic

z-a-f committed 4 years ago

225c38b7

Enabling intra-op parallelism (#26692)

dskhudia committed 4 years ago

8074526e

Suppressing hypothesis health check for qnnpack_add

z-a-f committed 4 years ago

06a866de

Handle uninitialized min/max values in histogram observer (#27151)

raghuramank100 committed 4 years ago

9b22a554

Unify quantized conv and linear tests (#26992)

supriyar committed 4 years ago

f8d3eac4

Uninitialize the accumulation buffer to save some overhead (#27005)

jianyuh committed 4 years ago

68b4d22d

Fuse module enhancements (#26457)

raghuramank100 committed 4 years ago

66b73b09

Add control for observers in Fake-quantize module (#27113)

raghuramank100 committed 4 years ago

32eb3b8d

Support for add relu functional module (#26612)

raghuramank100 committed 4 years ago

5a2a34cd

Default observer and fake-quant for backends (#26627)

raghuramank100 committed 4 years ago

a8083e18

Emulate weight and activation only quant with fake quant, numerics test (#26625)

raghuramank100 committed 4 years ago

bc3fb36e

Quantization aware training: Freeze batch norm support (#26624)

raghuramank100 committed 4 years ago

e0822f10

Per channel fake quant (#26623)

raghuramank100 committed 4 years ago

cbfd4e05

Newer Older