nihuini
|
b5b486fbfa
|
conv3x3s2 pack8 arm fp16sa neon assembly optimization
|
5 years ago |
Zhuo Zhang
|
418047661c
|
fix #1984 & fix cmake (#2000)
|
5 years ago |
ncnnnnn
|
e2557c1678
|
fix UNIT64_MAX not declared #2009 (#2010)
|
5 years ago |
nihui
|
5a9c99ce00
|
convdw5x5s1 pack8 arm neon fp16sa assembly optimization
|
5 years ago |
nihuini
|
e841ae73c6
|
fix arm fp16s feat output, fix #2003
|
5 years ago |
nihuini
|
8df3a02391
|
unroll 12 for conv1x1s1 and conv3x3s1 winograd pack8 arm fp16sa
|
5 years ago |
nihui
|
d8e9fc1443
|
conv3x3s1 conv3x3s2 pack1to8, padding pack8, relu pack8 arm neon fp16sa assmebly optimization
|
5 years ago |
kingdeviljin
|
ac9cbaca56
|
#1993 resize_bilinear_c4 fix (#1999)
|
5 years ago |
nihuini
|
b53f4072ce
|
convdw3x3s1 condw3x3s2 pack8 arm fp16sa
|
5 years ago |
nihuini
|
f6d808b090
|
crop pack8 arm fp16s, conv3x3s2 pack1to8 arm fp16sa intrinsic
|
5 years ago |
nihuini
|
bc05a71a7c
|
conv1x1s1 conv3x3s1 winograd arm fp16sa neon assembly optimization
|
5 years ago |
nihuini
|
5d5a3d1434
|
conv1x1s1 conv1x1s2 conv3x3s1 winograd pack8 arm fp16sa
|
5 years ago |
nihuini
|
20a0fc8628
|
packing honor thread count
|
5 years ago |
nihui
|
54e79a62d7
|
fix crash on non-arm82 build
|
5 years ago |
nihui
|
c173d51c9b
|
mish sigmoid swish tanh arm fp16s
|
5 years ago |
nihui
|
72a27d4776
|
utility wrapper for neon float32 bfloat16 conversion, deconvolution deconvolutiondepthwise arm fp16s fp16sa bf16s
|
5 years ago |
nihui
|
e644164873
|
reshape arm bf16s fp16s, flatten api
|
5 years ago |
nihui
|
aa68246dc7
|
more test coverage
|
5 years ago |
nihui
|
e7abc5fbd7
|
concat slice arm fp16sa pack8
|
5 years ago |
nihui
|
aa1a9e90c5
|
interp shufflechannel arm fp16sa pack8
|
5 years ago |
nihuini
|
c6d7525367
|
convolutiondepthwise arm fp16sa pack8
|
5 years ago |
nihuini
|
bc3822acc3
|
convolution flatten arm fp16sa pack8
|
5 years ago |
nihuini
|
dbb761b9a4
|
binaryop eltwise padding pooling arm fp16sa pack8
|
5 years ago |
nihuini
|
91d91ba556
|
hardsigmoid hardswish arm fp16s fp16sa
|
5 years ago |
nihuini
|
f23122bb3f
|
since fp16 storage option is on by default, upper-level function may pass fp32 storage with default option, guard with element bits checking
|
5 years ago |
nihuini
|
a18a9fd8c5
|
eltwise arm fp16s fp16sa
|
5 years ago |
nihuini
|
8385d81afa
|
interp arm fp16s fp16sa
|
5 years ago |
nihuini
|
5a4243e44e
|
binaryop arm fp16s
|
5 years ago |
nihuini
|
301abe657c
|
relu arm fp16s
|
5 years ago |
nihui
|
03c9ed11d2
|
pooling arm fp16s fp16sa
|
5 years ago |
nihui
|
71f86af8a6
|
fix non-arm82 ci
|
5 years ago |
nihui
|
9a2e2a6937
|
convert fp32 blobs for layers with fp16 storage support
|
5 years ago |
nihui
|
d4d501a7fe
|
fix innerproduct fp16sa
|
5 years ago |
nihui
|
e9c71a1ead
|
innerproduct arm fp16s fp16sa
|
5 years ago |
nihui
|
1a57600bd7
|
fix ci crash
|
5 years ago |
nihuini
|
11f5033249
|
convolutiondepthwise arm fp16s fp16sa
|
5 years ago |
nihuini
|
6ab284bc3a
|
convolution arm fp16s fp16sa
|
5 years ago |
nihuini
|
6d2c0e5683
|
flatten fp16s
|
5 years ago |
nihuini
|
47ae0c151a
|
some shared arm bf16s fp16s implementation
|
5 years ago |
zchrissirhcz
|
b80b84fda5
|
fix #1542; fix avx2 uint16_t including (#1968)
* fix #1542; fix avx2 uint16_t including
for #1542, it is for compatibility for opencv 2.x, such as on ubuntu 16.04 apt installed opencv
|
5 years ago |
nihui
|
1322ae40cb
|
update engine version
|
5 years ago |
nihui
|
308145254e
|
mask bf16 option in layer forward, disable gpu when bf16 enabled, fix #1962
|
5 years ago |
nihui
|
71dc13625f
|
disable bf16 storage for int8 inference
|
5 years ago |
nihuini
|
8700985540
|
yet another workaround for nexus6p gpu
|
5 years ago |
nihuini
|
bf279dcf17
|
workaround corrupted pipeline cache on old qcom adreno
|
5 years ago |
nihuini
|
4e4f0baa73
|
set openmp blocktime 20 for reducing power consumption, blocktime option
|
5 years ago |
nihui
|
21762e09e5
|
fix dilated convolution (#1956)
|
5 years ago |
nihuini
|
4d2d625432
|
fix avx2 build, second try, fix #1953
|
5 years ago |
nihuini
|
8b0890999a
|
fix avx2 build, fix #1953
|
5 years ago |
nihui
|
88367f4164
|
Ci enable mips msa (#1949)
|
5 years ago |