wget -c https://pjreddie.com/media/files/coco/trainvalno5k.part -O trainvalno5k.part
paste <(awk "{print \"$PWD\"}" <trainvalno5k.part) trainvalno5k.part | tr -d '\t' > trainvalno5k.txtwget -c https://pjreddie.com/media/files/coco/5k.part -O 5k.part
paste <(awk "{print \"$PWD\"}" <5k.part) 5k.part | tr -d '\t' > 5k.txtls -l coco/images/train2014
...
-rw-rw-r-- 1 jasonc jasonc 70995 Aug 16 2014 COCO_train2014_000000581921.jpg
-rw-rw-r-- 1 jasonc jasonc 77 Jul 12 18:06 COCO_train2014_000000581921.txt
...cat coco/images/train2014/COCO_train2014_000000581921.txt
...
0 0.425047 0.276405 0.334344 0.516838
31 0.462406 0.498700 0.064469 0.249859python3 COCO2YOLO/COCO2YOLO.py \
-j coco/images/annotations/instances_train2014.json \
-o coco/images/train2014ls coco/images/train2014
...
COCO_train2014_000000083060.jpg COCO_train2014_000000166137.txt COCO_train2014_000000250516.txt COCO_train2014_000000332960.jpg COCO_train2014_000000415119.jpg COCO_train2014_000000498082.txt COCO_train2014_000000581904.jpg
COCO_train2014_000000083060.txt COCO_train2014_000000166141.jpg COCO_train2014_000000250517.jpg COCO_train2014_000000332960.txt COCO_train2014_000000415119.txt COCO_train2014_000000498090.jpg COCO_train2014_000000581904.txt
COCO_train2014_000000083079.jpg COCO_train2014_000000166141.txt COCO_train2014_000000250517.txt COCO_train2014_000000332965.jpg COCO_train2014_000000415131.jpg COCO_train2014_000000498090.txt COCO_train2014_000000581906.jpg
COCO_train2014_000000083079.txt COCO_train2014_000000166163.jpg COCO_train2014_000000250518.jpg COCO_train2014_000000332965.txt COCO_train2014_000000415131.txt COCO_train2014_000000498091.jpg COCO_train2014_000000581906.txt
COCO_train2014_000000083085.jpg COCO_train2014_000000166163.txt COCO_train2014_000000250518.txt COCO_train2014_000000332976.jpg COCO_train2014_000000415146.jpg COCO_train2014_000000498091.txt COCO_train2014_000000581909.jpg
COCO_train2014_000000083085.txt COCO_train2014_000000166173.jpg COCO_train2014_000000250526.jpg COCO_train2014_000000332976.txt COCO_train2014_000000415146.txt COCO_train2014_000000498114.jpg COCO_train2014_000000581909.txt
COCO_train2014_000000083090.jpg COCO_train2014_000000166173.txt COCO_train2014_000000250526.txt COCO_train2014_000000333024.jpg COCO_train2014_000000415150.jpg COCO_train2014_000000498114.txt COCO_train2014_000000581921.jpg
COCO_train2014_000000083090.txt COCO_train2014_000000166179.jpg COCO_train2014_000000250533.jpg COCO_train2014_000000333024.txt COCO_train2014_000000415150.txt COCO_train2014_000000498125.jpg COCO_train2014_000000581921.txtsudo fc-cache -f -vdiff --git a/data/labels/make_labels.py b/data/labels/make_labels.py
index c8146f6..e1ffccf 100644
--- a/data/labels/make_labels.py
+++ b/data/labels/make_labels.py
@@ -2,22 +2,29 @@ import os
import string
import pipes
-font = 'futura-normal'
+font = '/usr/share/fonts/truetype/dongle/Futura_Medium.ttf'
def make_labels(s):
l = string.printable
for word in l:
if word == ' ':
- os.system('convert -fill black -background white -bordercolor white -font %s -pointsize %d label:"\ " 32_%d.png'%(font,s,s/12-1))
- if word == '@':
- os.system('convert -fill black -background white -bordercolor white -font %s -pointsize %d label:"\@" 64_%d.png'%(font,s,s/12-1))
+ err = os.system('convert-im6.q16 -fill black -background white -bordercolor white -font %s -pointsize %d label:"\ " 32_%d.png'%(font,s,s/12-1) )
+ if 0 != err:
+ print('ERR 0')
+ elif word == '@':
+ err = os.system('convert -fill black -background white -bordercolor white -font %s -pointsize %d label:"\@" 64_%d.png'%(font,s,s/12-1))
+ if 0 != err:
+ print('ERR 1')
elif word == '\\':
- os.system('convert -fill black -background white -bordercolor white -font %s -pointsize %d label:"\\\\\\\\" 92_%d.png'%(font,s,s/12-1))
+ err = os.system('convert -fill black -background white -bordercolor white -font %s -pointsize %d label:"\\\\\\\\" 92_%d.png'%(font,s,s/12-1))
+ if 0 != err:
+ print('ERR 2')
elif ord(word) in [9,10,11,12,13,14]:
pass
else:
- os.system("convert -fill black -background white -bordercolor white -font %s -pointsize %d label:%s \"%d_%d.png\""%(font,s,pipes.quote(word), ord(word),s/12-1))
+ err = os.system("convert -fill black -background white -bordercolor white -font %s -pointsize %d label:%s \"%d_%d.png\""%(font,s,pipes.quote(word), ord(word),s/12-1))
+ if 0 != err:
+ print('ERR 3')
for i in [12,24,36,48,60,72,84,96]:
make_labels(i)
-Patch the make_labels.py accordingly.
python3 make_labels.pyThe following png files for Labels shall be generated.
100_0.png 113_5.png 32_2.png 45_7.png 59_4.png 73_1.png 86_6.png
100_1.png 113_6.png 32_3.png 46_0.png 59_5.png 73_2.png 86_7.png
100_2.png 113_7.png 32_4.png 46_1.png 59_6.png 73_3.png 87_0.png
100_3.png 114_0.png 32_5.png 46_2.png 59_7.png 73_4.png 87_1.png
100_4.png 114_1.png 32_6.png 46_3.png 60_0.png 73_5.png 87_2.png
100_5.png 114_2.png 32_7.png 46_4.png 60_1.png 73_6.png 87_3.png
100_6.png 114_3.png 33_0.png 46_5.png 60_2.png 73_7.png 87_4.png
100_7.png 114_4.png 33_1.png 46_6.png 60_3.png 74_0.png 87_5.png
101_0.png 114_5.png 33_2.png 46_7.png 60_4.png 74_1.png 87_6.png
101_1.png 114_6.png 33_3.png 47_0.png 60_5.png 74_2.png 87_7.png
101_2.png 114_7.png 33_4.png 47_1.png 60_6.png 74_3.png 88_0.png
101_3.png 115_0.png 33_5.png 47_2.png 60_7.png 74_4.png 88_1.png
101_4.png 115_1.png 33_6.png 47_3.png 61_0.png 74_5.png 88_2.png
101_5.png 115_2.png 33_7.png 47_4.png 61_1.png 74_6.png 88_3.png
101_6.png 115_3.png 34_0.png 47_5.png 61_2.png 74_7.png 88_4.png
101_7.png 115_4.png 34_1.png 47_6.png 61_3.png 75_0.png 88_5.png
102_0.png 115_5.png 34_2.png 47_7.png 61_4.png 75_1.png 88_6.png
102_1.png 115_6.png 34_3.png 48_0.png 61_5.png 75_2.png 88_7.png
102_2.png 115_7.png 34_4.png 48_1.png 61_6.png 75_3.png 89_0.png
102_3.png 116_0.png 34_5.png 48_2.png 61_7.png 75_4.png 89_1.png
102_4.png 116_1.png 34_6.png 48_3.png 62_0.png 75_5.png 89_2.png
102_5.png 116_2.png 34_7.png 48_4.png 62_1.png 75_6.png 89_3.png
102_6.png 116_3.png 35_0.png 48_5.png 62_2.png 75_7.png 89_4.png
102_7.png 116_4.png 35_1.png 48_6.png 62_3.png 76_0.png 89_5.png
103_0.png 116_5.png 35_2.png 48_7.png 62_4.png 76_1.png 89_6.png
103_1.png 116_6.png 35_3.png 49_0.png 62_5.png 76_2.png 89_7.png
103_2.png 116_7.png 35_4.png 49_1.png 62_6.png 76_3.png 90_0.png
103_3.png 117_0.png 35_5.png 49_2.png 62_7.png 76_4.png 90_1.png
103_4.png 117_1.png 35_6.png 49_3.png 63_0.png 76_5.png 90_2.png
103_5.png 117_2.png 35_7.png 49_4.png 63_1.png 76_6.png 90_3.png
103_6.png 117_3.png 36_0.png 49_5.png 63_2.png 76_7.png 90_4.png
103_7.png 117_4.png 36_1.png 49_6.png 63_3.png 77_0.png 90_5.png
104_0.png 117_5.png 36_2.png 49_7.png 63_4.png 77_1.png 90_6.png
104_1.png 117_6.png 36_3.png 50_0.png 63_5.png 77_2.png 90_7.png
104_2.png 117_7.png 36_4.png 50_1.png 63_6.png 77_3.png 91_0.png
104_3.png 118_0.png 36_5.png 50_2.png 63_7.png 77_4.png 91_1.png
104_4.png 118_1.png 36_6.png 50_3.png 64_0.png 77_5.png 91_2.png
104_5.png 118_2.png 36_7.png 50_4.png 64_1.png 77_6.png 91_3.png
104_6.png 118_3.png 37_0.png 50_5.png 64_2.png 77_7.png 91_4.png
104_7.png 118_4.png 37_1.png 50_6.png 64_3.png 78_0.png 91_5.png
105_0.png 118_5.png 37_2.png 50_7.png 64_4.png 78_1.png 91_6.png
105_1.png 118_6.png 37_3.png 51_0.png 64_5.png 78_2.png 91_7.png
105_2.png 118_7.png 37_4.png 51_1.png 64_6.png 78_3.png 92_0.png
105_3.png 119_0.png 37_5.png 51_2.png 64_7.png 78_4.png 92_1.png
105_4.png 119_1.png 37_6.png 51_3.png 65_0.png 78_5.png 92_2.png
105_5.png 119_2.png 37_7.png 51_4.png 65_1.png 78_6.png 92_3.png
105_6.png 119_3.png 38_0.png 51_5.png 65_2.png 78_7.png 92_4.png
105_7.png 119_4.png 38_1.png 51_6.png 65_3.png 79_0.png 92_5.png
106_0.png 119_5.png 38_2.png 51_7.png 65_4.png 79_1.png 92_6.png
106_1.png 119_6.png 38_3.png 52_0.png 65_5.png 79_2.png 92_7.png
106_2.png 119_7.png 38_4.png 52_1.png 65_6.png 79_3.png 93_0.png
106_3.png 120_0.png 38_5.png 52_2.png 65_7.png 79_4.png 93_1.png
106_4.png 120_1.png 38_6.png 52_3.png 66_0.png 79_5.png 93_2.png
106_5.png 120_2.png 38_7.png 52_4.png 66_1.png 79_6.png 93_3.png
106_6.png 120_3.png 39_0.png 52_5.png 66_2.png 79_7.png 93_4.png
106_7.png 120_4.png 39_1.png 52_6.png 66_3.png 80_0.png 93_5.png
107_0.png 120_5.png 39_2.png 52_7.png 66_4.png 80_1.png 93_6.png
107_1.png 120_6.png 39_3.png 53_0.png 66_5.png 80_2.png 93_7.png
107_2.png 120_7.png 39_4.png 53_1.png 66_6.png 80_3.png 94_0.png
107_3.png 121_0.png 39_5.png 53_2.png 66_7.png 80_4.png 94_1.png
107_4.png 121_1.png 39_6.png 53_3.png 67_0.png 80_5.png 94_2.png
107_5.png 121_2.png 39_7.png 53_4.png 67_1.png 80_6.png 94_3.png
107_6.png 121_3.png 40_0.png 53_5.png 67_2.png 80_7.png 94_4.png
107_7.png 121_4.png 40_1.png 53_6.png 67_3.png 81_0.png 94_5.png
108_0.png 121_5.png 40_2.png 53_7.png 67_4.png 81_1.png 94_6.png
108_1.png 121_6.png 40_3.png 54_0.png 67_5.png 81_2.png 94_7.png
108_2.png 121_7.png 40_4.png 54_1.png 67_6.png 81_3.png 95_0.png
108_3.png 122_0.png 40_5.png 54_2.png 67_7.png 81_4.png 95_1.png
108_4.png 122_1.png 40_6.png 54_3.png 68_0.png 81_5.png 95_2.png
108_5.png 122_2.png 40_7.png 54_4.png 68_1.png 81_6.png 95_3.png
108_6.png 122_3.png 41_0.png 54_5.png 68_2.png 81_7.png 95_4.png
108_7.png 122_4.png 41_1.png 54_6.png 68_3.png 82_0.png 95_5.png
109_0.png 122_5.png 41_2.png 54_7.png 68_4.png 82_1.png 95_6.png
109_1.png 122_6.png 41_3.png 55_0.png 68_5.png 82_2.png 95_7.png
109_2.png 122_7.png 41_4.png 55_1.png 68_6.png 82_3.png 96_0.png
109_3.png 123_0.png 41_5.png 55_2.png 68_7.png 82_4.png 96_1.png
109_4.png 123_1.png 41_6.png 55_3.png 69_0.png 82_5.png 96_2.png
109_5.png 123_2.png 41_7.png 55_4.png 69_1.png 82_6.png 96_3.png
109_6.png 123_3.png 42_0.png 55_5.png 69_2.png 82_7.png 96_4.png
109_7.png 123_4.png 42_1.png 55_6.png 69_3.png 83_0.png 96_5.png
110_0.png 123_5.png 42_2.png 55_7.png 69_4.png 83_1.png 96_6.png
110_1.png 123_6.png 42_3.png 56_0.png 69_5.png 83_2.png 96_7.png
110_2.png 123_7.png 42_4.png 56_1.png 69_6.png 83_3.png 97_0.png
110_3.png 124_0.png 42_5.png 56_2.png 69_7.png 83_4.png 97_1.png
110_4.png 124_1.png 42_6.png 56_3.png 70_0.png 83_5.png 97_2.png
110_5.png 124_2.png 42_7.png 56_4.png 70_1.png 83_6.png 97_3.png
110_6.png 124_3.png 43_0.png 56_5.png 70_2.png 83_7.png 97_4.png
110_7.png 124_4.png 43_1.png 56_6.png 70_3.png 84_0.png 97_5.png
111_0.png 124_5.png 43_2.png 56_7.png 70_4.png 84_1.png 97_6.png
111_1.png 124_6.png 43_3.png 57_0.png 70_5.png 84_2.png 97_7.png
111_2.png 124_7.png 43_4.png 57_1.png 70_6.png 84_3.png 98_0.png
111_3.png 125_0.png 43_5.png 57_2.png 70_7.png 84_4.png 98_1.png
111_4.png 125_1.png 43_6.png 57_3.png 71_0.png 84_5.png 98_2.png
111_5.png 125_2.png 43_7.png 57_4.png 71_1.png 84_6.png 98_3.png
111_6.png 125_3.png 44_0.png 57_5.png 71_2.png 84_7.png 98_4.png
111_7.png 125_4.png 44_1.png 57_6.png 71_3.png 85_0.png 98_5.png
112_0.png 125_5.png 44_2.png 57_7.png 71_4.png 85_1.png 98_6.png
112_1.png 125_6.png 44_3.png 58_0.png 71_5.png 85_2.png 98_7.png
112_2.png 125_7.png 44_4.png 58_1.png 71_6.png 85_3.png 99_0.png
112_3.png 126_0.png 44_5.png 58_2.png 71_7.png 85_4.png 99_1.png
112_4.png 126_1.png 44_6.png 58_3.png 72_0.png 85_5.png 99_2.png
112_5.png 126_2.png 44_7.png 58_4.png 72_1.png 85_6.png 99_3.png
112_6.png 126_3.png 45_0.png 58_5.png 72_2.png 85_7.png 99_4.png
112_7.png 126_4.png 45_1.png 58_6.png 72_3.png 86_0.png 99_5.png
113_0.png 126_5.png 45_2.png 58_7.png 72_4.png 86_1.png 99_6.png
113_1.png 126_6.png 45_3.png 59_0.png 72_5.png 86_2.png 99_7.png
113_2.png 126_7.png 45_4.png 59_1.png 72_6.png 86_3.png
113_3.png 32_0.png 45_5.png 59_2.png 72_7.png 86_4.png
113_4.png 32_1.png 45_6.png 59_3.png 73_0.png 86_5.png../darknet detector train cfg/yolo-person.data cfg/yolo-person.cfg -gpus 0
../darknet detector test \
cfg/yolo-person.data \
cfg/yolo-person.cfg backup/yolo-person_final.weights \
pixmaps/people.jpg \
-thresh 0.40 \
-dont_show CUDA-version: 11070 (12010), cuDNN: 8.9.2, GPU count: 1
OpenCV version: 4.2.0
0 : compute_capability = 610, cudnn_half = 0, GPU: NVIDIA GeForce MX250
net.optimized_memory = 0
mini_batch = 1, batch = 1, time_steps = 1, train = 0
layer filters size/strd(dil) input output
0 Create CUDA-stream - 0
Create cudnn-handle 0
conv 8 3 x 3/ 2 160 x 160 x 1 -> 80 x 80 x 8 0.001 BF
1 conv 8 1 x 1/ 1 80 x 80 x 8 -> 80 x 80 x 8 0.001 BF
2 conv 8/ 8 3 x 3/ 1 80 x 80 x 8 -> 80 x 80 x 8 0.001 BF
3 conv 4 1 x 1/ 1 80 x 80 x 8 -> 80 x 80 x 4 0.000 BF
4 conv 8 1 x 1/ 1 80 x 80 x 4 -> 80 x 80 x 8 0.000 BF
5 conv 8/ 8 3 x 3/ 1 80 x 80 x 8 -> 80 x 80 x 8 0.001 BF
6 conv 4 1 x 1/ 1 80 x 80 x 8 -> 80 x 80 x 4 0.000 BF
7 dropout p = 0.150 25600 -> 25600
8 Shortcut Layer: 3, wt = 0, wn = 0, outputs: 80 x 80 x 4 0.000 BF
9 conv 24 1 x 1/ 1 80 x 80 x 4 -> 80 x 80 x 24 0.001 BF
10 conv 24/ 24 3 x 3/ 2 80 x 80 x 24 -> 40 x 40 x 24 0.001 BF
11 conv 8 1 x 1/ 1 40 x 40 x 24 -> 40 x 40 x 8 0.001 BF
12 conv 32 1 x 1/ 1 40 x 40 x 8 -> 40 x 40 x 32 0.001 BF
13 conv 32/ 32 3 x 3/ 1 40 x 40 x 32 -> 40 x 40 x 32 0.001 BF
14 conv 8 1 x 1/ 1 40 x 40 x 32 -> 40 x 40 x 8 0.001 BF
15 dropout p = 0.150 12800 -> 12800
16 Shortcut Layer: 11, wt = 0, wn = 0, outputs: 40 x 40 x 8 0.000 BF
17 conv 32 1 x 1/ 1 40 x 40 x 8 -> 40 x 40 x 32 0.001 BF
18 conv 32/ 32 3 x 3/ 1 40 x 40 x 32 -> 40 x 40 x 32 0.001 BF
19 conv 8 1 x 1/ 1 40 x 40 x 32 -> 40 x 40 x 8 0.001 BF
20 dropout p = 0.150 12800 -> 12800
21 Shortcut Layer: 16, wt = 0, wn = 0, outputs: 40 x 40 x 8 0.000 BF
22 conv 32 1 x 1/ 1 40 x 40 x 8 -> 40 x 40 x 32 0.001 BF
23 conv 32/ 32 3 x 3/ 2 40 x 40 x 32 -> 20 x 20 x 32 0.000 BF
24 conv 8 1 x 1/ 1 20 x 20 x 32 -> 20 x 20 x 8 0.000 BF
25 conv 48 1 x 1/ 1 20 x 20 x 8 -> 20 x 20 x 48 0.000 BF
26 conv 48/ 48 3 x 3/ 1 20 x 20 x 48 -> 20 x 20 x 48 0.000 BF
27 conv 8 1 x 1/ 1 20 x 20 x 48 -> 20 x 20 x 8 0.000 BF
28 dropout p = 0.150 3200 -> 3200
29 Shortcut Layer: 24, wt = 0, wn = 0, outputs: 20 x 20 x 8 0.000 BF
30 conv 48 1 x 1/ 1 20 x 20 x 8 -> 20 x 20 x 48 0.000 BF
31 conv 48/ 48 3 x 3/ 1 20 x 20 x 48 -> 20 x 20 x 48 0.000 BF
32 conv 8 1 x 1/ 1 20 x 20 x 48 -> 20 x 20 x 8 0.000 BF
33 dropout p = 0.150 3200 -> 3200
34 Shortcut Layer: 29, wt = 0, wn = 0, outputs: 20 x 20 x 8 0.000 BF
35 conv 48 1 x 1/ 1 20 x 20 x 8 -> 20 x 20 x 48 0.000 BF
36 conv 48/ 48 3 x 3/ 1 20 x 20 x 48 -> 20 x 20 x 48 0.000 BF
37 conv 16 1 x 1/ 1 20 x 20 x 48 -> 20 x 20 x 16 0.001 BF
38 conv 96 1 x 1/ 1 20 x 20 x 16 -> 20 x 20 x 96 0.001 BF
39 conv 96/ 96 3 x 3/ 1 20 x 20 x 96 -> 20 x 20 x 96 0.001 BF
40 conv 16 1 x 1/ 1 20 x 20 x 96 -> 20 x 20 x 16 0.001 BF
41 dropout p = 0.150 6400 -> 6400
42 Shortcut Layer: 37, wt = 0, wn = 0, outputs: 20 x 20 x 16 0.000 BF
43 conv 96 1 x 1/ 1 20 x 20 x 16 -> 20 x 20 x 96 0.001 BF
44 conv 96/ 96 3 x 3/ 1 20 x 20 x 96 -> 20 x 20 x 96 0.001 BF
45 conv 16 1 x 1/ 1 20 x 20 x 96 -> 20 x 20 x 16 0.001 BF
46 dropout p = 0.150 6400 -> 6400
47 Shortcut Layer: 42, wt = 0, wn = 0, outputs: 20 x 20 x 16 0.000 BF
48 conv 96 1 x 1/ 1 20 x 20 x 16 -> 20 x 20 x 96 0.001 BF
49 conv 96/ 96 3 x 3/ 1 20 x 20 x 96 -> 20 x 20 x 96 0.001 BF
50 conv 16 1 x 1/ 1 20 x 20 x 96 -> 20 x 20 x 16 0.001 BF
51 dropout p = 0.150 6400 -> 6400
52 Shortcut Layer: 47, wt = 0, wn = 0, outputs: 20 x 20 x 16 0.000 BF
53 conv 96 1 x 1/ 1 20 x 20 x 16 -> 20 x 20 x 96 0.001 BF
54 conv 96/ 96 3 x 3/ 1 20 x 20 x 96 -> 20 x 20 x 96 0.001 BF
55 conv 16 1 x 1/ 1 20 x 20 x 96 -> 20 x 20 x 16 0.001 BF
56 dropout p = 0.150 6400 -> 6400
57 Shortcut Layer: 52, wt = 0, wn = 0, outputs: 20 x 20 x 16 0.000 BF
58 conv 96 1 x 1/ 1 20 x 20 x 16 -> 20 x 20 x 96 0.001 BF
59 conv 96/ 96 3 x 3/ 2 20 x 20 x 96 -> 10 x 10 x 96 0.000 BF
60 conv 24 1 x 1/ 1 10 x 10 x 96 -> 10 x 10 x 24 0.000 BF
61 conv 136 1 x 1/ 1 10 x 10 x 24 -> 10 x 10 x 136 0.001 BF
62 conv 136/ 136 3 x 3/ 1 10 x 10 x 136 -> 10 x 10 x 136 0.000 BF
63 conv 24 1 x 1/ 1 10 x 10 x 136 -> 10 x 10 x 24 0.001 BF
64 dropout p = 0.150 2400 -> 2400
65 Shortcut Layer: 60, wt = 0, wn = 0, outputs: 10 x 10 x 24 0.000 BF
66 conv 136 1 x 1/ 1 10 x 10 x 24 -> 10 x 10 x 136 0.001 BF
67 conv 136/ 136 3 x 3/ 1 10 x 10 x 136 -> 10 x 10 x 136 0.000 BF
68 conv 24 1 x 1/ 1 10 x 10 x 136 -> 10 x 10 x 24 0.001 BF
69 dropout p = 0.150 2400 -> 2400
70 Shortcut Layer: 65, wt = 0, wn = 0, outputs: 10 x 10 x 24 0.000 BF
71 conv 136 1 x 1/ 1 10 x 10 x 24 -> 10 x 10 x 136 0.001 BF
72 conv 136/ 136 3 x 3/ 1 10 x 10 x 136 -> 10 x 10 x 136 0.000 BF
73 conv 24 1 x 1/ 1 10 x 10 x 136 -> 10 x 10 x 24 0.001 BF
74 dropout p = 0.150 2400 -> 2400
75 Shortcut Layer: 70, wt = 0, wn = 0, outputs: 10 x 10 x 24 0.000 BF
76 conv 136 1 x 1/ 1 10 x 10 x 24 -> 10 x 10 x 136 0.001 BF
77 conv 136/ 136 3 x 3/ 1 10 x 10 x 136 -> 10 x 10 x 136 0.000 BF
78 conv 24 1 x 1/ 1 10 x 10 x 136 -> 10 x 10 x 24 0.001 BF
79 dropout p = 0.150 2400 -> 2400
80 Shortcut Layer: 75, wt = 0, wn = 0, outputs: 10 x 10 x 24 0.000 BF
81 conv 136 1 x 1/ 1 10 x 10 x 24 -> 10 x 10 x 136 0.001 BF
82 conv 136/ 136 3 x 3/ 2 10 x 10 x 136 -> 5 x 5 x 136 0.000 BF
83 conv 48 1 x 1/ 1 5 x 5 x 136 -> 5 x 5 x 48 0.000 BF
84 conv 224 1 x 1/ 1 5 x 5 x 48 -> 5 x 5 x 224 0.001 BF
85 conv 224/ 224 3 x 3/ 1 5 x 5 x 224 -> 5 x 5 x 224 0.000 BF
86 conv 48 1 x 1/ 1 5 x 5 x 224 -> 5 x 5 x 48 0.001 BF
87 dropout p = 0.150 1200 -> 1200
88 Shortcut Layer: 83, wt = 0, wn = 0, outputs: 5 x 5 x 48 0.000 BF
89 conv 224 1 x 1/ 1 5 x 5 x 48 -> 5 x 5 x 224 0.001 BF
90 conv 224/ 224 3 x 3/ 1 5 x 5 x 224 -> 5 x 5 x 224 0.000 BF
91 conv 48 1 x 1/ 1 5 x 5 x 224 -> 5 x 5 x 48 0.001 BF
92 dropout p = 0.150 1200 -> 1200
93 Shortcut Layer: 88, wt = 0, wn = 0, outputs: 5 x 5 x 48 0.000 BF
94 conv 224 1 x 1/ 1 5 x 5 x 48 -> 5 x 5 x 224 0.001 BF
95 conv 224/ 224 3 x 3/ 1 5 x 5 x 224 -> 5 x 5 x 224 0.000 BF
96 conv 48 1 x 1/ 1 5 x 5 x 224 -> 5 x 5 x 48 0.001 BF
97 dropout p = 0.150 1200 -> 1200
98 Shortcut Layer: 93, wt = 0, wn = 0, outputs: 5 x 5 x 48 0.000 BF
99 conv 224 1 x 1/ 1 5 x 5 x 48 -> 5 x 5 x 224 0.001 BF
100 conv 224/ 224 3 x 3/ 1 5 x 5 x 224 -> 5 x 5 x 224 0.000 BF
101 conv 48 1 x 1/ 1 5 x 5 x 224 -> 5 x 5 x 48 0.001 BF
102 dropout p = 0.150 1200 -> 1200
103 Shortcut Layer: 98, wt = 0, wn = 0, outputs: 5 x 5 x 48 0.000 BF
104 conv 224 1 x 1/ 1 5 x 5 x 48 -> 5 x 5 x 224 0.001 BF
105 conv 224/ 224 3 x 3/ 1 5 x 5 x 224 -> 5 x 5 x 224 0.000 BF
106 conv 48 1 x 1/ 1 5 x 5 x 224 -> 5 x 5 x 48 0.001 BF
107 dropout p = 0.150 1200 -> 1200
108 Shortcut Layer: 103, wt = 0, wn = 0, outputs: 5 x 5 x 48 0.000 BF
109 max 3x 3/ 1 5 x 5 x 48 -> 5 x 5 x 48 0.000 BF
110 route 108 -> 5 x 5 x 48
111 max 5x 5/ 1 5 x 5 x 48 -> 5 x 5 x 48 0.000 BF
112 route 108 -> 5 x 5 x 48
113 max 9x 9/ 1 5 x 5 x 48 -> 5 x 5 x 48 0.000 BF
114 route 113 111 109 108 -> 5 x 5 x 192
115 conv 96 1 x 1/ 1 5 x 5 x 192 -> 5 x 5 x 96 0.001 BF
116 conv 96/ 96 5 x 5/ 1 5 x 5 x 96 -> 5 x 5 x 96 0.000 BF
117 conv 96 1 x 1/ 1 5 x 5 x 96 -> 5 x 5 x 96 0.000 BF
118 conv 96/ 96 5 x 5/ 1 5 x 5 x 96 -> 5 x 5 x 96 0.000 BF
119 conv 96 1 x 1/ 1 5 x 5 x 96 -> 5 x 5 x 96 0.000 BF
120 conv 18 1 x 1/ 1 5 x 5 x 96 -> 5 x 5 x 18 0.000 BF
121 yolo
[yolo] params: iou loss: ciou (4), iou_norm: 0.07, obj_norm: 1.00, cls_norm: 1.00, delta_norm: 1.00, scale_x_y: 1.00
nms_kind: greedynms (1), beta = 0.600000
122 route 115 -> 5 x 5 x 96
123 upsample 2x 5 x 5 x 96 -> 10 x 10 x 96
124 route 123 80 -> 10 x 10 x 120
125 conv 120/ 120 5 x 5/ 1 10 x 10 x 120 -> 10 x 10 x 120 0.001 BF
126 conv 120 1 x 1/ 1 10 x 10 x 120 -> 10 x 10 x 120 0.003 BF
127 conv 120/ 120 5 x 5/ 1 10 x 10 x 120 -> 10 x 10 x 120 0.001 BF
128 conv 120 1 x 1/ 1 10 x 10 x 120 -> 10 x 10 x 120 0.003 BF
129 conv 18 1 x 1/ 1 10 x 10 x 120 -> 10 x 10 x 18 0.000 BF
130 yolo
[yolo] params: iou loss: ciou (4), iou_norm: 0.07, obj_norm: 1.00, cls_norm: 1.00, delta_norm: 1.00, scale_x_y: 1.00
nms_kind: greedynms (1), beta = 0.600000
Total BFLOPS 0.054
avg_outputs = 15199
Allocate additional workspace_size = 0.04 MB
Loading weights from backup/yolo-person_last.weights...
seen 64, trained: 3568 K-images (55 Kilo-batches_64)
Done! Loaded 131 layers from weights-file
Detection layer: 121 - type = 28
Detection layer: 130 - type = 28
pixmaps/people.jpg: Predicted in 439.069000 milli-seconds.
person: 56%
person: 99%
person: 77%
person: 28%
person: 65%
person: 98%
person: 98%
person: 57%
person: 43%
person: 100%
person: 89%
person: 40%
person: 81%
person: 80%
person: 40%typing-extensions==3.10.0
python-dateutil==2.8.2
packaging==21.2
flatbuffers==23.1.21
requests==2.31.0
chardet==4.0.0
elastic-transport==8.0.0
google-auth==2.15.0
protobuf==3.20.3
urllib3==1.26.2
grpcio==1.48.2
testresources
numpy==1.23.5
setuptools
scipy
scikit-learn==0.20.3
opencv-python==4.2.0.32
opencv-contrib-python==4.2.0.32
tensorflow==2.13.0
tensorrt
keras_applications
tensorflow-model-optimization==0.5.0
tensorflow-addons
matplotlib
tqdm
pillow
mnn
Cython
pycocotools
keras2onnx
tf2onnx==0.4.2
onnx
onnxruntime
tfcoreml==1.1
sympy
imgaug
imagecorruptions
bokeh==2.4.0
tidecvpip3 install -r requirements.txtpython3 ../keras-YOLOv3-model-set/tools/model_converter/fastest_1.1_160/convert.py \
--config_path cfg/yolo-person.cfg \
--weights_path backup/yolo-person.weights \
--output_path backup/yolo-person.h5python3 ../keras-YOLOv3-model-set/tools/model_converter/fastest_1.1_160/post_train_quant_convert_demo.py \
--keras_model_file backup/yolo-person.h5 \
--annotation_file coco/trainvalno5k.txt \
--output_file backup/yolo-person.tfliteNote please, we use the trainvalno5k.txt annotation file for --annotation_file input parameter.
xxd -i backup/yolo-person.tflite > backup/yolo-person.cc