| .. |
|
activation_kernels.cu
|
Added CUDA-streams to Darknet-Yolo forward inference
|
7 years ago |
|
activation_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
activation_layer.h
|
idk
|
9 years ago |
|
activations.c
|
checkpoint
|
9 years ago |
|
activations.h
|
adding new tiny-yolo
|
9 years ago |
|
art.c
|
Supported OpenCV 3.0 and 2.4.13. Supported Windows and Linux.
|
8 years ago |
|
avgpool_layer.c
|
🔥 🔥 :dragonite:
|
9 years ago |
|
avgpool_layer.h
|
shortcut layers, msr networks
|
10 years ago |
|
avgpool_layer_kernels.cu
|
Fix for cuda 7.5
|
10 years ago |
|
batchnorm_layer.c
|
:vegan: :charizard:
|
9 years ago |
|
batchnorm_layer.h
|
stable
|
9 years ago |
|
blas.c
|
Fixed reorg_layer. Old reorg can be used as [reorg_old] in the cfg-file.
|
7 years ago |
|
blas.h
|
It takes into account the Difficult for calculating mAP for PascalVOC
|
7 years ago |
|
blas_kernels.cu
|
Added CUDA-streams to Darknet-Yolo forward inference
|
7 years ago |
|
box.c
|
Changed class to class_id
|
7 years ago |
|
box.h
|
CVPR Experiments
|
10 years ago |
|
captcha.c
|
better multigpu
|
9 years ago |
|
cifar.c
|
Changed class to class_id
|
7 years ago |
|
classifier.c
|
Changed class to class_id
|
7 years ago |
|
classifier.h
|
so much need to commit
|
9 years ago |
|
coco.c
|
Compile fix
|
7 years ago |
|
col2im.c
|
updates and things
|
9 years ago |
|
col2im.h
|
Detection is back, baby\!
|
10 years ago |
|
col2im_kernels.cu
|
updates and things
|
9 years ago |
|
compare.c
|
Changed class to class_id
|
7 years ago |
|
connected_layer.c
|
🔥 🔥 :dragonite:
|
9 years ago |
|
connected_layer.h
|
adding new tiny-yolo
|
9 years ago |
|
convolutional_kernels.cu
|
Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
|
7 years ago |
|
convolutional_layer.c
|
Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
|
7 years ago |
|
convolutional_layer.h
|
Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
|
7 years ago |
|
cost_layer.c
|
🔥 🔥 :dragonite:
|
9 years ago |
|
cost_layer.h
|
shortcut layers, msr networks
|
10 years ago |
|
cpu_gemm.c
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
crnn_layer.c
|
ADAM
|
9 years ago |
|
crnn_layer.h
|
stuff
|
9 years ago |
|
crop_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
crop_layer.h
|
lots of stuff
|
9 years ago |
|
crop_layer_kernels.cu
|
lots of stuff
|
9 years ago |
|
cuda.c
|
Tracking and Streaming fixed.
|
7 years ago |
|
cuda.h
|
If there is excessive GPU-RAM consumption by CUDNN then then do not use Workspace
|
7 years ago |
|
darknet.c
|
cheating a little here
|
9 years ago |
|
data.c
|
Changed class to class_id
|
7 years ago |
|
data.h
|
Fixed compile error
|
7 years ago |
|
deconvolutional_kernels.cu
|
good chance I didn't break anything
|
9 years ago |
|
deconvolutional_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
deconvolutional_layer.h
|
lots of stuff
|
9 years ago |
|
demo.c
|
Fixed saving video file
|
7 years ago |
|
demo.h
|
Compile fix
|
7 years ago |
|
detection_layer.c
|
It takes into account the Difficult for calculating mAP for PascalVOC
|
7 years ago |
|
detection_layer.h
|
So I have this new programming paradigm.......
|
9 years ago |
|
detector.c
|
Show avg-loss chart during Training, if compiled with OpenCV. Use -dont_show to disable.
|
7 years ago |
|
dice.c
|
better multigpu
|
9 years ago |
|
dropout_layer.c
|
🔥 🔥 :dragonite:
|
9 years ago |
|
dropout_layer.h
|
lots of stuff
|
9 years ago |
|
dropout_layer_kernels.cu
|
lots of stuff
|
9 years ago |
|
gemm.c
|
Added CUDA-streams to Darknet-Yolo forward inference
|
7 years ago |
|
gemm.h
|
go updates
|
9 years ago |
|
getopt.c
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
getopt.h
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
gettimeofday.c
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
gettimeofday.h
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
go.c
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
gru_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
gru_layer.h
|
So I have this new programming paradigm.......
|
9 years ago |
|
http_stream.cpp
|
Fix
|
7 years ago |
|
http_stream.h
|
Added param -http_port 8090 to show MJPEG-stream in the WebBrowser (Chrome/Firefox)
|
7 years ago |
|
im2col.c
|
updates and things
|
9 years ago |
|
im2col.h
|
Detection is back, baby\!
|
10 years ago |
|
im2col_kernels.cu
|
Added CUDA-streams to Darknet-Yolo forward inference
|
7 years ago |
|
image.c
|
Fixed saving video file
|
7 years ago |
|
image.h
|
Fixed classifier for AlexNet and Resnet50
|
8 years ago |
|
layer.c
|
Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
|
7 years ago |
|
layer.h
|
Fixed reorg_layer. Old reorg can be used as [reorg_old] in the cfg-file.
|
7 years ago |
|
list.c
|
Slowly refactoring and pushing to GPU
|
11 years ago |
|
list.h
|
Parsing, image loading, lots of stuff
|
12 years ago |
|
local_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
local_layer.h
|
shortcut layers, msr networks
|
10 years ago |
|
matrix.c
|
Changed class to class_id
|
7 years ago |
|
matrix.h
|
Go
|
9 years ago |
|
maxpool_layer.c
|
cpu batch norm works
|
9 years ago |
|
maxpool_layer.h
|
updates and things
|
9 years ago |
|
maxpool_layer_kernels.cu
|
Added CUDA-streams to Darknet-Yolo forward inference
|
7 years ago |
|
network.c
|
Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
|
7 years ago |
|
network.h
|
Detection and Tracking work well
|
7 years ago |
|
network_kernels.cu
|
Added support for Tensor Cores CC >= 7.0 (V100). For FP16/32 (mixed precision) define CUDNN_HALF should be used.
|
7 years ago |
|
nightmare.c
|
stuff
|
9 years ago |
|
normalization_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
normalization_layer.h
|
shortcut layers, msr networks
|
10 years ago |
|
option_list.c
|
new font strategy
|
9 years ago |
|
option_list.h
|
new font strategy
|
9 years ago |
|
parser.c
|
Fixed reorg_layer. Old reorg can be used as [reorg_old] in the cfg-file.
|
7 years ago |
|
parser.h
|
Fixed for Linux: detection for batch > 1 and 0x0d at command line
|
8 years ago |
|
region_layer.c
|
Changed class to class_id
|
7 years ago |
|
region_layer.h
|
Parameter (max) in the cfg-file has an effect on the Yolo-training
|
7 years ago |
|
reorg_layer.c
|
Added: calc_anchors
|
7 years ago |
|
reorg_layer.h
|
stuff and things
|
9 years ago |
|
reorg_old_layer.c
|
Fixed reorg_layer. Old reorg can be used as [reorg_old] in the cfg-file.
|
7 years ago |
|
reorg_old_layer.h
|
Fixed reorg_layer. Old reorg can be used as [reorg_old] in the cfg-file.
|
7 years ago |
|
rnn.c
|
better multigpu
|
9 years ago |
|
rnn_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
rnn_layer.h
|
So I have this new programming paradigm.......
|
9 years ago |
|
rnn_vid.c
|
Supported OpenCV 3.0 and 2.4.13. Supported Windows and Linux.
|
8 years ago |
|
route_layer.c
|
🔥 🔥 yolo v2 🔥 🔥
|
9 years ago |
|
route_layer.h
|
hey
|
9 years ago |
|
shortcut_layer.c
|
So I have this new programming paradigm.......
|
9 years ago |
|
shortcut_layer.h
|
shortcut layers, msr networks
|
10 years ago |
|
softmax_layer.c
|
It takes into account the Difficult for calculating mAP for PascalVOC
|
7 years ago |
|
softmax_layer.h
|
lots of stuff
|
9 years ago |
|
stb_image.h
|
Fixed STB
|
8 years ago |
|
stb_image_write.h
|
Fixed STB
|
8 years ago |
|
super.c
|
better multigpu
|
9 years ago |
|
swag.c
|
better multigpu
|
9 years ago |
|
tag.c
|
better multigpu
|
9 years ago |
|
tree.c
|
hey
|
9 years ago |
|
tree.h
|
hey
|
9 years ago |
|
unistd.h
|
Sources fixed - Yolo v2 can be used on Windows (MSVS 2015 + CUDA 8.0)
|
9 years ago |
|
utils.c
|
Replaced '\r' and '\n' by 0x0d and 0x0a
|
7 years ago |
|
utils.h
|
More usage rand_s()
|
8 years ago |
|
voxel.c
|
Supported OpenCV 3.0 and 2.4.13. Supported Windows and Linux.
|
8 years ago |
|
writing.c
|
better multigpu
|
9 years ago |
|
yolo.c
|
Compile fix
|
7 years ago |
|
yolo_console_dll.cpp
|
Minor fix
|
7 years ago |
|
yolo_v2_class.cpp
|
Changed definition location to adhere to compatability promises
|
7 years ago |
|
yolo_v2_class.hpp
|
Fixed Tracking using SparsePyrLKOpticalFlow on CPU
|
7 years ago |
AlexeyAB
f0bc634a74