mirror of
https://github.com/pjreddie/darknet.git
synced 2023-08-10 21:13:14 +03:00
resnet that works
This commit is contained in:
Joseph Redmon
parent
1b001a7f58
commit
62b781af4d
4
Makefile
4
Makefile
@ -4,11 +4,11 @@ OPENCV=0
|
||||
OPENMP=0
|
||||
DEBUG=0
|
||||
|
||||
ARCH= -gencode arch=compute_20,code=[sm_20,sm_21] \
|
||||
-gencode arch=compute_30,code=sm_30 \
|
||||
ARCH= -gencode arch=compute_30,code=sm_30 \
|
||||
-gencode arch=compute_35,code=sm_35 \
|
||||
-gencode arch=compute_50,code=[sm_50,compute_50] \
|
||||
-gencode arch=compute_52,code=[sm_52,compute_52]
|
||||
# -gencode arch=compute_20,code=[sm_20,sm_21] \ This one is deprecated?
|
||||
|
||||
# This is what I use, uncomment if you know your arch and want to specify
|
||||
# ARCH= -gencode arch=compute_52,code=compute_52
|
||||
|
||||
366
cfg/msr_34.cfg
366
cfg/msr_34.cfg
@ -1,366 +0,0 @@
|
||||
[net]
|
||||
batch=128
|
||||
subdivisions=1
|
||||
height=256
|
||||
width=256
|
||||
channels=3
|
||||
momentum=0.9
|
||||
decay=0.0005
|
||||
|
||||
learning_rate=0.1
|
||||
policy=poly
|
||||
power=4
|
||||
max_batches=500000
|
||||
|
||||
[crop]
|
||||
crop_height=224
|
||||
crop_width=224
|
||||
flip=1
|
||||
saturation=1
|
||||
exposure=1
|
||||
angle=0
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=7
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[maxpool]
|
||||
size=3
|
||||
stride=2
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
|
||||
[avgpool]
|
||||
|
||||
[connected]
|
||||
output=1000
|
||||
activation=leaky
|
||||
|
||||
[softmax]
|
||||
groups=1
|
||||
|
||||
[cost]
|
||||
type=sse
|
||||
|
||||
558
cfg/msr_50.cfg
558
cfg/msr_50.cfg
@ -1,558 +0,0 @@
|
||||
[net]
|
||||
batch=128
|
||||
subdivisions=8
|
||||
height=256
|
||||
width=256
|
||||
channels=3
|
||||
momentum=0.9
|
||||
decay=0.0001
|
||||
|
||||
learning_rate=0.05
|
||||
policy=poly
|
||||
power=4
|
||||
max_batches=500000
|
||||
|
||||
|
||||
|
||||
[crop]
|
||||
crop_height=224
|
||||
crop_width=224
|
||||
flip=1
|
||||
saturation=1
|
||||
exposure=1
|
||||
angle=0
|
||||
|
||||
##### Conv 1 #####
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=7
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[maxpool]
|
||||
size=3
|
||||
stride=2
|
||||
|
||||
|
||||
##### Conv 2_x #####
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=256
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
|
||||
##### Conv 3_x #####
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=2
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=512
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
|
||||
##### Conv 4_x #####
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=1024
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=2
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=1024
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=1024
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=1024
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=1024
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=1024
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=1024
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
|
||||
##### Conv 5_x #####
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=2
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=2048
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=2
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=2048
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=2048
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=2048
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
activation=leaky
|
||||
|
||||
[avgpool]
|
||||
|
||||
[connected]
|
||||
output=1000
|
||||
activation=leaky
|
||||
|
||||
[softmax]
|
||||
groups=1
|
||||
|
||||
[cost]
|
||||
type=sse
|
||||
|
||||
@ -1,26 +1,30 @@
|
||||
[net]
|
||||
batch=128
|
||||
subdivisions=8
|
||||
# Training
|
||||
# batch=128
|
||||
# subdivisions=8
|
||||
|
||||
# Testing
|
||||
batch=1
|
||||
subdivisions=1
|
||||
|
||||
height=256
|
||||
width=256
|
||||
max_crop=448
|
||||
channels=3
|
||||
momentum=0.9
|
||||
decay=0.0001
|
||||
decay=0.0005
|
||||
|
||||
burn_in=1000
|
||||
learning_rate=0.1
|
||||
policy=poly
|
||||
power=4
|
||||
max_batches=500000
|
||||
max_batches=1600000
|
||||
|
||||
[crop]
|
||||
crop_height=224
|
||||
crop_width=224
|
||||
flip=1
|
||||
saturation=1
|
||||
exposure=1
|
||||
angle=0
|
||||
|
||||
##### Conv 1 #####
|
||||
angle=7
|
||||
hue=.1
|
||||
saturation=.75
|
||||
exposure=.75
|
||||
aspect=.75
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
@ -31,13 +35,9 @@ pad=1
|
||||
activation=leaky
|
||||
|
||||
[maxpool]
|
||||
size=3
|
||||
size=2
|
||||
stride=2
|
||||
|
||||
|
||||
##### Conv 2_x #####
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=64
|
||||
@ -62,19 +62,8 @@ stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=256
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
@ -102,8 +91,7 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
@ -131,13 +119,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
##### Conv 3_x #####
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -162,23 +146,10 @@ stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=2
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=512
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -204,11 +175,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -234,11 +203,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -264,11 +231,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -294,11 +259,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -324,11 +287,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -354,11 +315,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=128
|
||||
@ -384,14 +343,11 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
|
||||
##### Conv 4_x #####
|
||||
|
||||
# Conv 4
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -416,23 +372,10 @@ stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=2
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=1024
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -458,11 +401,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -488,11 +429,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -518,11 +457,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -548,11 +485,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -578,11 +513,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -608,11 +541,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -638,11 +569,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -668,11 +597,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -698,11 +625,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -728,11 +653,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -758,11 +681,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -788,11 +709,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -818,11 +737,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -848,11 +765,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -878,11 +793,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -908,11 +821,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -938,11 +849,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -968,11 +877,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -998,11 +905,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1028,11 +933,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1058,11 +961,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1088,11 +989,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1118,11 +1017,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1148,11 +1045,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1178,11 +1073,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1208,11 +1101,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1238,11 +1129,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1268,11 +1157,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1298,11 +1185,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1328,11 +1213,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1358,11 +1241,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1388,11 +1269,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1418,11 +1297,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1448,11 +1325,9 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=256
|
||||
@ -1478,13 +1353,10 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
##### Conv 5_x #####
|
||||
|
||||
#Conv 5
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
@ -1509,50 +1381,8 @@ stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
|
||||
[route]
|
||||
layers=-4
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
size=1
|
||||
stride=2
|
||||
pad=1
|
||||
activation=linear
|
||||
filters=2048
|
||||
|
||||
[shortcut]
|
||||
from = -3
|
||||
activation=leaky
|
||||
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=2048
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
@ -1580,16 +1410,51 @@ pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from = -4
|
||||
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=512
|
||||
size=3
|
||||
stride=1
|
||||
pad=1
|
||||
activation=leaky
|
||||
|
||||
[convolutional]
|
||||
batch_normalize=1
|
||||
filters=2048
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[shortcut]
|
||||
from=-4
|
||||
activation=leaky
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
|
||||
[convolutional]
|
||||
filters=1000
|
||||
size=1
|
||||
stride=1
|
||||
pad=1
|
||||
activation=linear
|
||||
|
||||
[avgpool]
|
||||
|
||||
[connected]
|
||||
output=1000
|
||||
activation=leaky
|
||||
|
||||
[softmax]
|
||||
groups=1
|
||||
|
||||
@ -58,7 +58,7 @@ void train_classifier(char *datacfg, char *cfgfile, char *weightfile, int *gpus,
|
||||
load_args args = {0};
|
||||
args.w = net.w;
|
||||
args.h = net.h;
|
||||
args.threads = 64;
|
||||
args.threads = 32;
|
||||
args.hierarchy = net.hierarchy;
|
||||
|
||||
args.min = net.min_crop;
|
||||
@ -123,6 +123,7 @@ void train_classifier(char *datacfg, char *cfgfile, char *weightfile, int *gpus,
|
||||
char buff[256];
|
||||
sprintf(buff, "%s/%s.weights", backup_directory, base);
|
||||
save_weights(net, buff);
|
||||
pthread_join(load_thread, 0);
|
||||
|
||||
free_network(net);
|
||||
free_ptrs((void**)labels, classes);
|
||||
|
||||
@ -83,27 +83,8 @@ void average(int argc, char *argv[])
|
||||
save_weights(sum, outfile);
|
||||
}
|
||||
|
||||
void speed(char *cfgfile, int tics)
|
||||
long numops(network net)
|
||||
{
|
||||
if (tics == 0) tics = 1000;
|
||||
network net = parse_network_cfg(cfgfile);
|
||||
set_batch_network(&net, 1);
|
||||
int i;
|
||||
double time=what_time_is_it_now();
|
||||
image im = make_image(net.w, net.h, net.c*net.batch);
|
||||
for(i = 0; i < tics; ++i){
|
||||
network_predict(net, im.data);
|
||||
}
|
||||
double t = what_time_is_it_now() - time;
|
||||
printf("\n%d evals, %f Seconds\n", tics, t);
|
||||
printf("Speed: %f sec/eval\n", t/tics);
|
||||
printf("Speed: %f Hz\n", tics/t);
|
||||
}
|
||||
|
||||
void operations(char *cfgfile)
|
||||
{
|
||||
gpu_index = -1;
|
||||
network net = parse_network_cfg(cfgfile);
|
||||
int i;
|
||||
long ops = 0;
|
||||
for(i = 0; i < net.n; ++i){
|
||||
@ -134,6 +115,34 @@ void operations(char *cfgfile)
|
||||
ops += 2l * l.wo->inputs * l.wo->outputs;
|
||||
}
|
||||
}
|
||||
return ops;
|
||||
}
|
||||
|
||||
void speed(char *cfgfile, int tics)
|
||||
{
|
||||
if (tics == 0) tics = 1000;
|
||||
network net = parse_network_cfg(cfgfile);
|
||||
set_batch_network(&net, 1);
|
||||
int i;
|
||||
double time=what_time_is_it_now();
|
||||
image im = make_image(net.w, net.h, net.c*net.batch);
|
||||
for(i = 0; i < tics; ++i){
|
||||
network_predict(net, im.data);
|
||||
}
|
||||
double t = what_time_is_it_now() - time;
|
||||
long ops = numops(net);
|
||||
printf("\n%d evals, %f Seconds\n", tics, t);
|
||||
printf("Floating Point Operations: %.2f Bn\n", (float)ops/1000000000.);
|
||||
printf("FLOPS: %.2f Bn\n", (float)ops/1000000000.*tics/t);
|
||||
printf("Speed: %f sec/eval\n", t/tics);
|
||||
printf("Speed: %f Hz\n", tics/t);
|
||||
}
|
||||
|
||||
void operations(char *cfgfile)
|
||||
{
|
||||
gpu_index = -1;
|
||||
network net = parse_network_cfg(cfgfile);
|
||||
long ops = numops(net);
|
||||
printf("Floating Point Operations: %ld\n", ops);
|
||||
printf("Floating Point Operations: %.2f Bn\n", (float)ops/1000000000.);
|
||||
}
|
||||
|
||||
@ -52,10 +52,10 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
|
||||
args.d = &buffer;
|
||||
args.type = DETECTION_DATA;
|
||||
//args.type = INSTANCE_DATA;
|
||||
args.threads = 8;
|
||||
args.threads = 64;
|
||||
|
||||
pthread_t load_thread = load_data(args);
|
||||
clock_t time;
|
||||
double time;
|
||||
int count = 0;
|
||||
//while(i*imgs < N*120){
|
||||
while(get_current_batch(net) < net.max_batches){
|
||||
@ -78,7 +78,7 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
|
||||
}
|
||||
net = nets[0];
|
||||
}
|
||||
time=clock();
|
||||
time=what_time_is_it_now();
|
||||
pthread_join(load_thread, 0);
|
||||
train = buffer;
|
||||
load_thread = load_data(args);
|
||||
@ -107,9 +107,9 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
|
||||
}
|
||||
*/
|
||||
|
||||
printf("Loaded: %lf seconds\n", sec(clock()-time));
|
||||
printf("Loaded: %lf seconds\n", what_time_is_it_now()-time);
|
||||
|
||||
time=clock();
|
||||
time=what_time_is_it_now();
|
||||
float loss = 0;
|
||||
#ifdef GPU
|
||||
if(ngpus == 1){
|
||||
@ -124,7 +124,7 @@ void train_detector(char *datacfg, char *cfgfile, char *weightfile, int *gpus, i
|
||||
avg_loss = avg_loss*.9 + loss*.1;
|
||||
|
||||
i = get_current_batch(net);
|
||||
printf("%ld: %f, %f avg, %f rate, %lf seconds, %d images\n", get_current_batch(net), loss, avg_loss, get_current_rate(net), sec(clock()-time), i*imgs);
|
||||
printf("%ld: %f, %f avg, %f rate, %lf seconds, %d images\n", get_current_batch(net), loss, avg_loss, get_current_rate(net), what_time_is_it_now()-time, i*imgs);
|
||||
if(i%100==0){
|
||||
#ifdef GPU
|
||||
if(ngpus != 1) sync_nets(nets, ngpus, 0);
|
||||
@ -313,7 +313,7 @@ void validate_detector_flip(char *datacfg, char *cfgfile, char *weightfile, char
|
||||
args.resized = &buf_resized[t];
|
||||
thr[t] = load_data_in_thread(args);
|
||||
}
|
||||
time_t start = time(0);
|
||||
double start = what_time_is_it_now();
|
||||
for(i = nthreads; i < m+nthreads; i += nthreads){
|
||||
fprintf(stderr, "%d\n", i);
|
||||
for(t = 0; t < nthreads && i+t-nthreads < m; ++t){
|
||||
@ -359,7 +359,7 @@ void validate_detector_flip(char *datacfg, char *cfgfile, char *weightfile, char
|
||||
fprintf(fp, "\n]\n");
|
||||
fclose(fp);
|
||||
}
|
||||
fprintf(stderr, "Total Detection Time: %f Seconds\n", (double)(time(0) - start));
|
||||
fprintf(stderr, "Total Detection Time: %f Seconds\n", what_time_is_it_now() - start);
|
||||
}
|
||||
|
||||
|
||||
@ -447,7 +447,7 @@ void validate_detector(char *datacfg, char *cfgfile, char *weightfile, char *out
|
||||
args.resized = &buf_resized[t];
|
||||
thr[t] = load_data_in_thread(args);
|
||||
}
|
||||
time_t start = time(0);
|
||||
double start = what_time_is_it_now();
|
||||
for(i = nthreads; i < m+nthreads; i += nthreads){
|
||||
fprintf(stderr, "%d\n", i);
|
||||
for(t = 0; t < nthreads && i+t-nthreads < m; ++t){
|
||||
@ -490,7 +490,7 @@ void validate_detector(char *datacfg, char *cfgfile, char *weightfile, char *out
|
||||
fprintf(fp, "\n]\n");
|
||||
fclose(fp);
|
||||
}
|
||||
fprintf(stderr, "Total Detection Time: %f Seconds\n", (double)(time(0) - start));
|
||||
fprintf(stderr, "Total Detection Time: %f Seconds\n", what_time_is_it_now() - start);
|
||||
}
|
||||
|
||||
void validate_detector_recall(char *cfgfile, char *weightfile)
|
||||
|
||||
@ -52,6 +52,7 @@ char **read_tokens(char *filename, size_t *read)
|
||||
return d;
|
||||
}
|
||||
|
||||
|
||||
float_pair get_rnn_token_data(int *tokens, size_t *offsets, int characters, size_t len, int batch, int steps)
|
||||
{
|
||||
float *x = calloc(batch * steps * characters, sizeof(float));
|
||||
@ -78,6 +79,35 @@ float_pair get_rnn_token_data(int *tokens, size_t *offsets, int characters, size
|
||||
return p;
|
||||
}
|
||||
|
||||
float_pair get_seq2seq_data(char **source, char **dest, int n, int characters, size_t len, int batch, int steps)
|
||||
{
|
||||
int i,j;
|
||||
float *x = calloc(batch * steps * characters, sizeof(float));
|
||||
float *y = calloc(batch * steps * characters, sizeof(float));
|
||||
for(i = 0; i < batch; ++i){
|
||||
int index = rand()%n;
|
||||
for(j = 0; j < steps; ++j){
|
||||
unsigned char curr = source[index][j];
|
||||
unsigned char next = dest[index][j];
|
||||
|
||||
x[(j*batch + i)*characters + curr] = 1;
|
||||
y[(j*batch + i)*characters + next] = 1;
|
||||
|
||||
if(curr > 255 || curr <= 0 || next > 255 || next <= 0){
|
||||
/*text[(index+j+2)%len] = 0;
|
||||
printf("%ld %d %d %d %d\n", index, j, len, (int)text[index+j], (int)text[index+j+1]);
|
||||
printf("%s", text+index);
|
||||
*/
|
||||
error("Bad char");
|
||||
}
|
||||
}
|
||||
}
|
||||
float_pair p;
|
||||
p.x = x;
|
||||
p.y = y;
|
||||
return p;
|
||||
}
|
||||
|
||||
float_pair get_rnn_data(unsigned char *text, size_t *offsets, int characters, size_t len, int batch, int steps)
|
||||
{
|
||||
float *x = calloc(batch * steps * characters, sizeof(float));
|
||||
|
||||
@ -211,7 +211,7 @@ void demo_segmenter(char *datacfg, char *cfgfile, char *weightfile, int cam_inde
|
||||
image in = get_image_from_stream(cap);
|
||||
image in_s = letterbox_image(in, net.w, net.h);
|
||||
|
||||
float *predictions = network_predict(net, in_s.data);
|
||||
network_predict(net, in_s.data);
|
||||
|
||||
printf("\033[2J");
|
||||
printf("\033[1;1H");
|
||||
|
||||
@ -10,8 +10,8 @@ extern "C" {
|
||||
|
||||
__device__ float lhtan_activate_kernel(float x)
|
||||
{
|
||||
if(x < 0) return .001*x;
|
||||
if(x > 1) return .001*(x-1) + 1;
|
||||
if(x < 0) return .001f*x;
|
||||
if(x > 1) return .001f*(x-1.f) + 1.f;
|
||||
return x;
|
||||
}
|
||||
__device__ float lhtan_gradient_kernel(float x)
|
||||
@ -27,25 +27,25 @@ __device__ float hardtan_activate_kernel(float x)
|
||||
return x;
|
||||
}
|
||||
__device__ float linear_activate_kernel(float x){return x;}
|
||||
__device__ float logistic_activate_kernel(float x){return 1./(1. + exp(-x));}
|
||||
__device__ float loggy_activate_kernel(float x){return 2./(1. + exp(-x)) - 1;}
|
||||
__device__ float logistic_activate_kernel(float x){return 1.f/(1.f + expf(-x));}
|
||||
__device__ float loggy_activate_kernel(float x){return 2.f/(1.f + expf(-x)) - 1;}
|
||||
__device__ float relu_activate_kernel(float x){return x*(x>0);}
|
||||
__device__ float elu_activate_kernel(float x){return (x >= 0)*x + (x < 0)*(exp(x)-1);}
|
||||
__device__ float relie_activate_kernel(float x){return (x>0) ? x : .01*x;}
|
||||
__device__ float ramp_activate_kernel(float x){return x*(x>0)+.1*x;}
|
||||
__device__ float leaky_activate_kernel(float x){return (x>0) ? x : .1*x;}
|
||||
__device__ float tanh_activate_kernel(float x){return (2/(1 + exp(-2*x)) - 1);}
|
||||
__device__ float elu_activate_kernel(float x){return (x >= 0)*x + (x < 0)*(expf(x)-1);}
|
||||
__device__ float relie_activate_kernel(float x){return (x>0) ? x : .01f*x;}
|
||||
__device__ float ramp_activate_kernel(float x){return x*(x>0)+.1f*x;}
|
||||
__device__ float leaky_activate_kernel(float x){return (x>0) ? x : .1f*x;}
|
||||
__device__ float tanh_activate_kernel(float x){return (2.f/(1 + expf(-2*x)) - 1);}
|
||||
__device__ float plse_activate_kernel(float x)
|
||||
{
|
||||
if(x < -4) return .01 * (x + 4);
|
||||
if(x > 4) return .01 * (x - 4) + 1;
|
||||
return .125*x + .5;
|
||||
if(x < -4) return .01f * (x + 4);
|
||||
if(x > 4) return .01f * (x - 4) + 1;
|
||||
return .125f*x + .5f;
|
||||
}
|
||||
__device__ float stair_activate_kernel(float x)
|
||||
{
|
||||
int n = floor(x);
|
||||
if (n%2 == 0) return floor(x/2.);
|
||||
else return (x - n) + floor(x/2.);
|
||||
int n = floorf(x);
|
||||
if (n%2 == 0) return floorf(x/2);
|
||||
else return (x - n) + floorf(x/2);
|
||||
}
|
||||
|
||||
|
||||
@ -58,19 +58,19 @@ __device__ float linear_gradient_kernel(float x){return 1;}
|
||||
__device__ float logistic_gradient_kernel(float x){return (1-x)*x;}
|
||||
__device__ float loggy_gradient_kernel(float x)
|
||||
{
|
||||
float y = (x+1.)/2.;
|
||||
float y = (x+1)/2;
|
||||
return 2*(1-y)*y;
|
||||
}
|
||||
__device__ float relu_gradient_kernel(float x){return (x>0);}
|
||||
__device__ float elu_gradient_kernel(float x){return (x >= 0) + (x < 0)*(x + 1);}
|
||||
__device__ float relie_gradient_kernel(float x){return (x>0) ? 1 : .01;}
|
||||
__device__ float ramp_gradient_kernel(float x){return (x>0)+.1;}
|
||||
__device__ float leaky_gradient_kernel(float x){return (x>0) ? 1 : .1;}
|
||||
__device__ float relie_gradient_kernel(float x){return (x>0) ? 1 : .01f;}
|
||||
__device__ float ramp_gradient_kernel(float x){return (x>0)+.1f;}
|
||||
__device__ float leaky_gradient_kernel(float x){return (x>0) ? 1 : .1f;}
|
||||
__device__ float tanh_gradient_kernel(float x){return 1-x*x;}
|
||||
__device__ float plse_gradient_kernel(float x){return (x < 0 || x > 1) ? .01 : .125;}
|
||||
__device__ float plse_gradient_kernel(float x){return (x < 0 || x > 1) ? .01f : .125f;}
|
||||
__device__ float stair_gradient_kernel(float x)
|
||||
{
|
||||
if (floor(x) == x) return 0;
|
||||
if (floorf(x) == x) return 0;
|
||||
return 1;
|
||||
}
|
||||
|
||||
|
||||
@ -165,7 +165,7 @@ __global__ void adam_kernel(int N, float *x, float *m, float *v, float B1, float
|
||||
int index = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
|
||||
if (index >= N) return;
|
||||
|
||||
x[index] = x[index] + (rate * sqrt(1.-pow(B2, t)) / (1.-pow(B1, t)) * m[index] / (sqrt(v[index]) + eps));
|
||||
x[index] = x[index] + (rate * sqrtf(1.f-powf(B2, t)) / (1.f-powf(B1, t)) * m[index] / (sqrtf(v[index]) + eps));
|
||||
}
|
||||
|
||||
extern "C" void adam_gpu(int n, float *x, float *m, float *v, float B1, float B2, float rate, float eps, int t)
|
||||
@ -194,7 +194,7 @@ __global__ void normalize_kernel(int N, float *x, float *mean, float *variance,
|
||||
if (index >= N) return;
|
||||
int f = (index/spatial)%filters;
|
||||
|
||||
x[index] = (x[index] - mean[f])/(sqrt(variance[f] + .00001f));
|
||||
x[index] = (x[index] - mean[f])/(sqrtf(variance[f] + .00001f));
|
||||
}
|
||||
|
||||
__global__ void normalize_delta_kernel(int N, float *x, float *mean, float *variance, float *mean_delta, float *variance_delta, int batch, int filters, int spatial, float *delta)
|
||||
@ -203,7 +203,7 @@ __global__ void normalize_delta_kernel(int N, float *x, float *mean, float *vari
|
||||
if (index >= N) return;
|
||||
int f = (index/spatial)%filters;
|
||||
|
||||
delta[index] = delta[index] * 1./(sqrt(variance[f] + .00001f)) + variance_delta[f] * 2. * (x[index] - mean[f]) / (spatial * batch) + mean_delta[f]/(spatial*batch);
|
||||
delta[index] = delta[index] * 1.f/(sqrtf(variance[f] + .00001f)) + variance_delta[f] * 2.f * (x[index] - mean[f]) / (spatial * batch) + mean_delta[f]/(spatial*batch);
|
||||
}
|
||||
|
||||
extern "C" void normalize_delta_gpu(float *x, float *mean, float *variance, float *mean_delta, float *variance_delta, int batch, int filters, int spatial, float *delta)
|
||||
@ -225,7 +225,7 @@ __global__ void variance_delta_kernel(float *x, float *delta, float *mean, floa
|
||||
variance_delta[i] += delta[index]*(x[index] - mean[i]);
|
||||
}
|
||||
}
|
||||
variance_delta[i] *= -.5 * pow(variance[i] + .00001f, (float)(-3./2.));
|
||||
variance_delta[i] *= -.5f * powf(variance[i] + .00001f, (float)(-3.f/2.f));
|
||||
}
|
||||
|
||||
__global__ void accumulate_kernel(float *x, int n, int groups, float *sum)
|
||||
@ -264,7 +264,7 @@ __global__ void fast_mean_delta_kernel(float *delta, float *variance, int batch,
|
||||
for(i = 0; i < threads; ++i){
|
||||
mean_delta[filter] += local[i];
|
||||
}
|
||||
mean_delta[filter] *= (-1./sqrt(variance[filter] + .00001f));
|
||||
mean_delta[filter] *= (-1.f/sqrtf(variance[filter] + .00001f));
|
||||
}
|
||||
}
|
||||
|
||||
@ -294,7 +294,7 @@ __global__ void fast_variance_delta_kernel(float *x, float *delta, float *mean,
|
||||
for(i = 0; i < threads; ++i){
|
||||
variance_delta[filter] += local[i];
|
||||
}
|
||||
variance_delta[filter] *= -.5 * pow(variance[filter] + .00001f, (float)(-3./2.));
|
||||
variance_delta[filter] *= -.5f * powf(variance[filter] + .00001f, (float)(-3.f/2.f));
|
||||
}
|
||||
}
|
||||
|
||||
@ -311,7 +311,7 @@ __global__ void mean_delta_kernel(float *delta, float *variance, int batch, int
|
||||
mean_delta[i] += delta[index];
|
||||
}
|
||||
}
|
||||
mean_delta[i] *= (-1./sqrt(variance[i] + .00001f));
|
||||
mean_delta[i] *= (-1.f/sqrtf(variance[i] + .00001f));
|
||||
}
|
||||
|
||||
extern "C" void mean_delta_gpu(float *delta, float *variance, int batch, int filters, int spatial, float *mean_delta)
|
||||
@ -334,7 +334,7 @@ extern "C" void fast_variance_delta_gpu(float *x, float *delta, float *mean, flo
|
||||
|
||||
__global__ void mean_kernel(float *x, int batch, int filters, int spatial, float *mean)
|
||||
{
|
||||
float scale = 1./(batch * spatial);
|
||||
float scale = 1.f/(batch * spatial);
|
||||
int i = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
|
||||
if (i >= filters) return;
|
||||
int j,k;
|
||||
@ -350,7 +350,7 @@ __global__ void mean_kernel(float *x, int batch, int filters, int spatial, floa
|
||||
|
||||
__global__ void variance_kernel(float *x, float *mean, int batch, int filters, int spatial, float *variance)
|
||||
{
|
||||
float scale = 1./(batch * spatial - 1);
|
||||
float scale = 1.f/(batch * spatial - 1);
|
||||
int j,k;
|
||||
int i = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
|
||||
if (i >= filters) return;
|
||||
@ -358,7 +358,7 @@ __global__ void variance_kernel(float *x, float *mean, int batch, int filters, i
|
||||
for(j = 0; j < batch; ++j){
|
||||
for(k = 0; k < spatial; ++k){
|
||||
int index = j*filters*spatial + i*spatial + k;
|
||||
variance[i] += pow((x[index] - mean[i]), 2);
|
||||
variance[i] += powf((x[index] - mean[i]), 2);
|
||||
}
|
||||
}
|
||||
variance[i] *= scale;
|
||||
@ -516,7 +516,7 @@ __global__ void fast_variance_kernel(float *x, float *mean, int batch, int filt
|
||||
for(i = 0; i < spatial; i += threads){
|
||||
int index = j*spatial*filters + filter*spatial + i + id;
|
||||
|
||||
local[id] += (i+id < spatial) ? pow((x[index] - mean[filter]), 2) : 0;
|
||||
local[id] += (i+id < spatial) ? powf((x[index] - mean[filter]), 2) : 0;
|
||||
}
|
||||
}
|
||||
|
||||
@ -716,7 +716,7 @@ __global__ void smooth_l1_kernel(int n, float *pred, float *truth, float *delta,
|
||||
int i = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
|
||||
if(i < n){
|
||||
float diff = truth[i] - pred[i];
|
||||
float abs_val = abs(diff);
|
||||
float abs_val = fabsf(diff);
|
||||
if(abs_val < 1) {
|
||||
error[i] = diff * diff;
|
||||
delta[i] = diff;
|
||||
@ -864,7 +864,7 @@ __device__ void softmax_device(float *input, int n, float temp, int stride, floa
|
||||
largest = (val>largest) ? val : largest;
|
||||
}
|
||||
for(i = 0; i < n; ++i){
|
||||
float e = exp(input[i*stride]/temp - largest/temp);
|
||||
float e = expf(input[i*stride]/temp - largest/temp);
|
||||
sum += e;
|
||||
output[i*stride] = e;
|
||||
}
|
||||
|
||||
@ -33,7 +33,7 @@ __global__ void binarize_input_kernel(float *input, int n, int size, float *bina
|
||||
int i = 0;
|
||||
float mean = 0;
|
||||
for(i = 0; i < n; ++i){
|
||||
mean += abs(input[i*size + s]);
|
||||
mean += fabsf(input[i*size + s]);
|
||||
}
|
||||
mean = mean / n;
|
||||
for(i = 0; i < n; ++i){
|
||||
@ -55,7 +55,7 @@ __global__ void binarize_weights_kernel(float *weights, int n, int size, float *
|
||||
int i = 0;
|
||||
float mean = 0;
|
||||
for(i = 0; i < size; ++i){
|
||||
mean += abs(weights[f*size + i]);
|
||||
mean += fabsf(weights[f*size + i]);
|
||||
}
|
||||
mean = mean / size;
|
||||
for(i = 0; i < size; ++i){
|
||||
@ -139,8 +139,8 @@ __global__ void smooth_kernel(float *x, int n, int w, int h, int c, int size, fl
|
||||
id /= c;
|
||||
int b = id;
|
||||
|
||||
int w_offset = -(size/2.);
|
||||
int h_offset = -(size/2.);
|
||||
int w_offset = -(size/2.f);
|
||||
int h_offset = -(size/2.f);
|
||||
|
||||
int out_index = j + w*(i + h*(k + c*b));
|
||||
int l, m;
|
||||
|
||||
@ -113,9 +113,9 @@ __global__ void levels_image_kernel(float *image, float *rand, int batch, int w,
|
||||
float r3 = rand[8*id + 3];
|
||||
|
||||
saturation = r0*(saturation - 1) + 1;
|
||||
saturation = (r1 > .5) ? 1./saturation : saturation;
|
||||
saturation = (r1 > .5f) ? 1.f/saturation : saturation;
|
||||
exposure = r2*(exposure - 1) + 1;
|
||||
exposure = (r3 > .5) ? 1./exposure : exposure;
|
||||
exposure = (r3 > .5f) ? 1.f/exposure : exposure;
|
||||
|
||||
size_t offset = id * h * w * 3;
|
||||
image += offset;
|
||||
@ -131,9 +131,9 @@ __global__ void levels_image_kernel(float *image, float *rand, int batch, int w,
|
||||
} else {
|
||||
shift = 0;
|
||||
}
|
||||
image[x + w*(y + h*0)] = rgb.x*scale + translate + (rshift - .5)*shift;
|
||||
image[x + w*(y + h*1)] = rgb.y*scale + translate + (gshift - .5)*shift;
|
||||
image[x + w*(y + h*2)] = rgb.z*scale + translate + (bshift - .5)*shift;
|
||||
image[x + w*(y + h*0)] = rgb.x*scale + translate + (rshift - .5f)*shift;
|
||||
image[x + w*(y + h*1)] = rgb.y*scale + translate + (gshift - .5f)*shift;
|
||||
image[x + w*(y + h*2)] = rgb.z*scale + translate + (bshift - .5f)*shift;
|
||||
}
|
||||
|
||||
__global__ void forward_crop_layer_kernel(float *input, float *rand, int size, int c, int h, int w, int crop_height, int crop_width, int train, int flip, float angle, float *output)
|
||||
@ -141,8 +141,8 @@ __global__ void forward_crop_layer_kernel(float *input, float *rand, int size, i
|
||||
int id = (blockIdx.x + blockIdx.y*gridDim.x) * blockDim.x + threadIdx.x;
|
||||
if(id >= size) return;
|
||||
|
||||
float cx = w/2.;
|
||||
float cy = h/2.;
|
||||
float cx = w/2.f;
|
||||
float cy = h/2.f;
|
||||
|
||||
int count = id;
|
||||
int j = id % crop_width;
|
||||
@ -160,11 +160,11 @@ __global__ void forward_crop_layer_kernel(float *input, float *rand, int size, i
|
||||
|
||||
float dw = (w - crop_width)*r4;
|
||||
float dh = (h - crop_height)*r5;
|
||||
flip = (flip && (r6 > .5));
|
||||
flip = (flip && (r6 > .5f));
|
||||
angle = 2*angle*r7 - angle;
|
||||
if(!train){
|
||||
dw = (w - crop_width)/2.;
|
||||
dh = (h - crop_height)/2.;
|
||||
dw = (w - crop_width)/2.f;
|
||||
dh = (h - crop_height)/2.f;
|
||||
flip = 0;
|
||||
angle = 0;
|
||||
}
|
||||
@ -174,8 +174,8 @@ __global__ void forward_crop_layer_kernel(float *input, float *rand, int size, i
|
||||
float x = (flip) ? w - dw - j - 1 : j + dw;
|
||||
float y = i + dh;
|
||||
|
||||
float rx = cos(angle)*(x-cx) - sin(angle)*(y-cy) + cx;
|
||||
float ry = sin(angle)*(x-cx) + cos(angle)*(y-cy) + cy;
|
||||
float rx = cosf(angle)*(x-cx) - sinf(angle)*(y-cy) + cx;
|
||||
float ry = sinf(angle)*(x-cx) + cosf(angle)*(y-cy) + cy;
|
||||
|
||||
output[count] = bilinear_interpolate_kernel(input, w, h, rx, ry, k);
|
||||
}
|
||||
@ -184,7 +184,7 @@ extern "C" void forward_crop_layer_gpu(crop_layer layer, network net)
|
||||
{
|
||||
cuda_random(layer.rand_gpu, layer.batch*8);
|
||||
|
||||
float radians = layer.angle*3.14159265/180.;
|
||||
float radians = layer.angle*3.14159265f/180.f;
|
||||
|
||||
float scale = 2;
|
||||
float translate = -1;
|
||||
|
||||
@ -137,14 +137,18 @@ matrix load_image_augment_paths(char **paths, int n, int min, int max, int size,
|
||||
|
||||
box_label *read_boxes(char *filename, int *n)
|
||||
{
|
||||
box_label *boxes = calloc(1, sizeof(box_label));
|
||||
FILE *file = fopen(filename, "r");
|
||||
if(!file) file_error(filename);
|
||||
float x, y, h, w;
|
||||
int id;
|
||||
int count = 0;
|
||||
int size = 64;
|
||||
box_label *boxes = calloc(size, sizeof(box_label));
|
||||
while(fscanf(file, "%d %f %f %f %f", &id, &x, &y, &w, &h) == 5){
|
||||
boxes = realloc(boxes, (count+1)*sizeof(box_label));
|
||||
if(count == size) {
|
||||
size = size * 2;
|
||||
boxes = realloc(boxes, size*sizeof(box_label));
|
||||
}
|
||||
boxes[count].id = id;
|
||||
boxes[count].x = x;
|
||||
boxes[count].y = y;
|
||||
@ -976,6 +980,7 @@ data load_data_detection(int n, char **paths, int m, int w, int h, int boxes, in
|
||||
place_image(orig, nw, nh, dx, dy, sized);
|
||||
|
||||
random_distort_image(sized, hue, saturation, exposure);
|
||||
|
||||
int flip = rand()%2;
|
||||
if(flip) flip_image(sized);
|
||||
d.X.vals[i] = sized.data;
|
||||
|
||||
@ -20,10 +20,6 @@ static int demo_classes;
|
||||
static float **probs;
|
||||
static box *boxes;
|
||||
static network net;
|
||||
static network net2;
|
||||
static float **probs2;
|
||||
static box *boxes2;
|
||||
static float **predictions2;
|
||||
static image buff [3];
|
||||
static image buff_letter[3];
|
||||
static int buff_index = 0;
|
||||
|
||||
90
src/image.c
90
src/image.c
@ -44,6 +44,51 @@ image mask_to_rgb(image mask)
|
||||
return im;
|
||||
}
|
||||
|
||||
static float get_pixel(image m, int x, int y, int c)
|
||||
{
|
||||
assert(x < m.w && y < m.h && c < m.c);
|
||||
return m.data[c*m.h*m.w + y*m.w + x];
|
||||
}
|
||||
static float get_pixel_extend(image m, int x, int y, int c)
|
||||
{
|
||||
if(x < 0 || x >= m.w || y < 0 || y >= m.h) return 0;
|
||||
/*
|
||||
if(x < 0) x = 0;
|
||||
if(x >= m.w) x = m.w-1;
|
||||
if(y < 0) y = 0;
|
||||
if(y >= m.h) y = m.h-1;
|
||||
*/
|
||||
if(c < 0 || c >= m.c) return 0;
|
||||
return get_pixel(m, x, y, c);
|
||||
}
|
||||
static void set_pixel(image m, int x, int y, int c, float val)
|
||||
{
|
||||
if (x < 0 || y < 0 || c < 0 || x >= m.w || y >= m.h || c >= m.c) return;
|
||||
assert(x < m.w && y < m.h && c < m.c);
|
||||
m.data[c*m.h*m.w + y*m.w + x] = val;
|
||||
}
|
||||
static void add_pixel(image m, int x, int y, int c, float val)
|
||||
{
|
||||
assert(x < m.w && y < m.h && c < m.c);
|
||||
m.data[c*m.h*m.w + y*m.w + x] += val;
|
||||
}
|
||||
|
||||
static float bilinear_interpolate(image im, float x, float y, int c)
|
||||
{
|
||||
int ix = (int) floorf(x);
|
||||
int iy = (int) floorf(y);
|
||||
|
||||
float dx = x - ix;
|
||||
float dy = y - iy;
|
||||
|
||||
float val = (1-dy) * (1-dx) * get_pixel_extend(im, ix, iy, c) +
|
||||
dy * (1-dx) * get_pixel_extend(im, ix, iy+1, c) +
|
||||
(1-dy) * dx * get_pixel_extend(im, ix+1, iy, c) +
|
||||
dy * dx * get_pixel_extend(im, ix+1, iy+1, c);
|
||||
return val;
|
||||
}
|
||||
|
||||
|
||||
void composite_image(image source, image dest, int dx, int dy)
|
||||
{
|
||||
int x,y,k;
|
||||
@ -1255,21 +1300,6 @@ void saturate_exposure_image(image im, float sat, float exposure)
|
||||
constrain_image(im);
|
||||
}
|
||||
|
||||
float bilinear_interpolate(image im, float x, float y, int c)
|
||||
{
|
||||
int ix = (int) floorf(x);
|
||||
int iy = (int) floorf(y);
|
||||
|
||||
float dx = x - ix;
|
||||
float dy = y - iy;
|
||||
|
||||
float val = (1-dy) * (1-dx) * get_pixel_extend(im, ix, iy, c) +
|
||||
dy * (1-dx) * get_pixel_extend(im, ix, iy+1, c) +
|
||||
(1-dy) * dx * get_pixel_extend(im, ix+1, iy, c) +
|
||||
dy * dx * get_pixel_extend(im, ix+1, iy+1, c);
|
||||
return val;
|
||||
}
|
||||
|
||||
image resize_image(image im, int w, int h)
|
||||
{
|
||||
image resized = make_image(w, h, im.c);
|
||||
@ -1419,36 +1449,6 @@ image get_image_layer(image m, int l)
|
||||
}
|
||||
return out;
|
||||
}
|
||||
|
||||
float get_pixel(image m, int x, int y, int c)
|
||||
{
|
||||
assert(x < m.w && y < m.h && c < m.c);
|
||||
return m.data[c*m.h*m.w + y*m.w + x];
|
||||
}
|
||||
float get_pixel_extend(image m, int x, int y, int c)
|
||||
{
|
||||
if(x < 0 || x >= m.w || y < 0 || y >= m.h) return 0;
|
||||
/*
|
||||
if(x < 0) x = 0;
|
||||
if(x >= m.w) x = m.w-1;
|
||||
if(y < 0) y = 0;
|
||||
if(y >= m.h) y = m.h-1;
|
||||
*/
|
||||
if(c < 0 || c >= m.c) return 0;
|
||||
return get_pixel(m, x, y, c);
|
||||
}
|
||||
void set_pixel(image m, int x, int y, int c, float val)
|
||||
{
|
||||
if (x < 0 || y < 0 || c < 0 || x >= m.w || y >= m.h || c >= m.c) return;
|
||||
assert(x < m.w && y < m.h && c < m.c);
|
||||
m.data[c*m.h*m.w + y*m.w + x] = val;
|
||||
}
|
||||
void add_pixel(image m, int x, int y, int c, float val)
|
||||
{
|
||||
assert(x < m.w && y < m.h && c < m.c);
|
||||
m.data[c*m.h*m.w + y*m.w + x] += val;
|
||||
}
|
||||
|
||||
void print_image(image m)
|
||||
{
|
||||
int i, j, k;
|
||||
|
||||
@ -60,12 +60,6 @@ void print_image(image m);
|
||||
image make_empty_image(int w, int h, int c);
|
||||
void copy_image_into(image src, image dest);
|
||||
|
||||
float get_pixel(image m, int x, int y, int c);
|
||||
float get_pixel_extend(image m, int x, int y, int c);
|
||||
void set_pixel(image m, int x, int y, int c, float val);
|
||||
void add_pixel(image m, int x, int y, int c, float val);
|
||||
float bilinear_interpolate(image im, float x, float y, int c);
|
||||
|
||||
image get_image_layer(image m, int l);
|
||||
|
||||
#endif
|
||||
|
||||
@ -5,7 +5,14 @@
|
||||
#include "darknet.h"
|
||||
#include "list.h"
|
||||
|
||||
#define TWO_PI 6.2831853071795864769252866
|
||||
#define TIME(a) \
|
||||
do { \
|
||||
double start = what_time_is_it_now(); \
|
||||
a; \
|
||||
printf("%s took: %f seconds\n", #a, what_time_is_it_now() - start); \
|
||||
} while (0)
|
||||
|
||||
#define TWO_PI 6.2831853071795864769252866f
|
||||
|
||||
double what_time_is_it_now();
|
||||
void shuffle(void *arr, size_t n, size_t size);
|
||||
|
||||
Loading…
Reference in New Issue
Block a user