nihuini
b6741f1053
bump vulkan sdk version
6 years ago
tpoisonooo
d706b7073a
add FAQ-ncnn-protobuf-problem.zh.md doc ( #1552 )
6 years ago
nihui
8ffcb2f973
add conv1x1 conv3x3 pack4 bf16s neon kernel
6 years ago
songqun
f5d6aba720
test_tanh.cpp: Add unittest of tanh ( #1578 )
6 years ago
nihui
44eb28fadc
fix cast arm packing test
6 years ago
nihui
52ce59e672
fix build with requant option on
6 years ago
nihui
f214883203
cast between float32 and bfloat16
6 years ago
nihuini
0b3d0ddb8c
ci build avx2 variants
6 years ago
nihuini
06668872e0
ci build nostdio nostring variants
6 years ago
nihuini
648ef3fdee
reuse vkallocator in test
6 years ago
nihuini
5ce8dbdb79
add how-to-build.md
6 years ago
nihui
4c5bdd7036
update faq
6 years ago
nihui
839c4c4e34
smaller test size
6 years ago
nihui
10fbce7bff
arm64 coverage ( #1571 )
* arm64 coverage
* second try
6 years ago
nihui
d227c3f446
add codecov badge
6 years ago
nihui
39d26c8373
test-coverage action ( #1569 )
* coverage
* wip
* use relative path
* enable gpu test
* restore jobs
* codecov
6 years ago
Cai Shanli
0281e51fe5
named all enum types ( #1570 )
6 years ago
nihui
26d7e379f7
sync-wiki action ( #1568 )
* Create sync-wiki.yml
* Update sync-wiki.yml
6 years ago
kalcohol
94444f09b6
light up the building status matrix ( #1564 )
* delete ccpp.yml
* add build.yml
* split all workflows
* delete single file defined workflows
* light up the building status matrix
* use short icons
* add actions to ios 32bit cpu position
* rename ios workflow file's name
* rename ios workflow name
6 years ago
kalcohol
06e129d259
add skip cooling down option ( #1566 )
* add option for cooling down
* add option for cooling down
add the option usage
add benchmark of RTX2080
* add benchmark of rtx2060
6 years ago
nihui
9929d52885
less duplicated code for crop layer, slice axes starts from 0
6 years ago
nihui
61ae6e865e
fix vkallocator flush invalidate size
6 years ago
nihui
5cba7518a1
add efficient roi resize rotate doc
6 years ago
nihui
abe65593fa
add ncnn opencv conversion doc
6 years ago
nihuini
be3f978770
fix onnx opset11 Pad conversion, improve Resize conversion, fix #1556
6 years ago
nihui
57bedd59fa
fix build without neon
6 years ago
nihui
ed810928ae
add ncnn-load-model doc
6 years ago
nihui
7c97142524
old qcom adreno driver seems to have the same bug as mali does
6 years ago
nihui
82fa27967e
priorbox shader fix
6 years ago
zhiliu6
2606d2a2c9
Fix AVX convolution problem. ( #1549 )
6 years ago
nihui
bdc6681d44
replace nullptr to 0, fix build on old toolchain, fix #1541
6 years ago
zhiliu6
01bfdb0ce5
Optimize FP32/FP16 conversion with AVX intrinsic. ( #1545 )
* optimize yolov3 output extraction speed.
* Optimize FP32/FP16 conversion with AVX intrinsics.
6 years ago
nihui
cd8e0d8045
fix crop arm without packing, fix #1546
6 years ago
nihui
362100936d
fix scale pack4
6 years ago
nihui
999da7158f
old glslang reject -Os option, as optimizing for size does not make a big difference, drop it for now, fix #1544
6 years ago
nihui
61956069ae
fix pack8 condition
6 years ago
nihui
9821be7a42
create pack8 pipeline only if shader_pack8 is enabled
6 years ago
nihui
90cc9c1a50
fix build
6 years ago
nihui
bbaa4dcce2
compile fp16pa, optimize shader for size, enable implicit fp16 arithmetic for qcom855 and qcom855plus
6 years ago
nihui
c089ddb3c2
fix build with opencv 2.4
6 years ago
nihui
719d9f48ae
im2col wrt sgemm convolution option
6 years ago
nihui
30c2954160
fuse onnx pixelshuffle
6 years ago
nihui
f1996411b6
fix instancenorm fp16p
6 years ago
nihui
4ca0c3c4ab
fix concat fp16p
6 years ago
nihui
759d653663
fix various fp16p issues, fix #1540
6 years ago
nihui
7ae585f217
shape hint is elemsize aware
6 years ago
nihui
f4b1760a38
pixelshuffle vulkan, pixelshuffle test
6 years ago
nihui
bc524b6da8
pixelshuffle
6 years ago
Cai Shanli
0ccd2da219
named enum ncnn::Mat::PixelType ( #1539 )
for pyncnn python wrapper
6 years ago
nihui
4fead31bea
some test case changes
6 years ago