nihuini
4a9e452cb5
fix build on msvc, first try
7 years ago
xinyu391
666ec2820e
add Net::load_param_mem(), ParamDict::load_param_mem() load plain text paramters from memory buf. ( #587 )
7 years ago
nihui
5d04a3a45c
layer holds bottom blob scale, depthwise convolution read group scales
8 years ago
nihui
354b95256c
bump param version, backward compatible
8 years ago
nihuini
2bc504925e
fix int8_scales from multiple blobs, fix #512
8 years ago
nihui
a169cec363
core int8 inference, quantize and dequantize, net using flag, caffe2ncnn reads int8 scale table
8 years ago
nihui
9706cd1447
implement ncnn blob/workspace allocator, fine-grained per-layer openmp threads control, fix #469
8 years ago
nihuini
a341e7465c
reject to load model with empty network, fix #392
8 years ago
Hyungsuk Yoon
354f515596
scanf with width specifier need (width + 1) spaces
8 years ago
nihui
6c4c810fda
decouple modelbin of different input types, simplify timestamp function
8 years ago
nihui
d68eb4cd15
wrap benchmark gettimeofday
8 years ago
Linghan Cheung
811b6ba1b6
print benchmark information for every layer, especially for CONVOLUTION ( #241 )
* print benchmark information for every layer, especially for CONVOLUTION
* print benchmark information for every layer, especially for CONVOLUTION, for cross-platform.
* move the function implementation to cpp file to avoid multiple definitions
8 years ago
nihui
10b86c2af5
create layer from type name
8 years ago
nihui
7a43c45e80
remove deprecated code
8 years ago
nihui
a181d25098
new model load api, fix #215
8 years ago
nihuini
b84ba31c23
enable light mode by default
8 years ago
nihui
1e2265dd99
new param load api
8 years ago
nihui
87bad3420e
improve custom layer register
8 years ago
nihuini
8eba811e4b
move layer type enum to separate header
8 years ago
tedder59
a26914d6c8
注册用户网络层小bug
注册用户网络层时,返回-1表示不存在同名的网络层。
9 years ago
nihuini
b7db8be4f6
add ncnn source qwq
9 years ago