Bienqq
/
car-detection-bayes
2
1
Fork 0
Code Issues Pull Requests Releases Wiki Activity

updates

This commit is contained in:
Glenn Jocher 2020-01-04 11:36:36 -08:00
parent efe3c319b5
commit d197c0be75
2 changed files with 9 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
utils/gcp.sh
View File

@ -3,10 +3,10 @@
# New VM
rm -rf sample_data yolov3
git clone https://github.com/ultralytics/yolov3
git clone https://github.com/NVIDIA/apex && cd apex && pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" . --user && cd .. && rm -rf apex
#git clone https://github.com/NVIDIA/apex && cd apex && pip install -v --no-cache-dir --global-option="--cpp_ext" --global-option="--cuda_ext" . --user && cd .. && rm -rf apex
sudo conda install -yc conda-forge scikit-image pycocotools
python3 -c "from yolov3.utils.google_utils import gdrive_download; gdrive_download('193Zp_ye-3qXMonR1nZj3YyxMtQkMy50k','coco2014.zip')"
# python3 -c "from yolov3.utils.google_utils import gdrive_download; gdrive_download('1WQT6SOktSe8Uw6r10-2JhbEhMY5DJaph','coco2017.zip')"
# python3 -c "from yolov3.utils.google_utils import gdrive_download; gdrive_download('193Zp_ye-3qXMonR1nZj3YyxMtQkMy50k','coco2014.zip')"
python3 -c "from yolov3.utils.google_utils import gdrive_download; gdrive_download('1WQT6SOktSe8Uw6r10-2JhbEhMY5DJaph','coco2017.zip')"
sudo shutdown
# Re-clone
@ -38,7 +38,7 @@ python3 detect.py
python3 test.py --save-json
# Evolve
t=ultralytics/yolov3:v179
t=ultralytics/yolov3:v189
sudo docker kill $(sudo docker ps -a -q --filter ancestor=$t)
for i in 0
do
@ -234,7 +234,10 @@ t=ultralytics/yolov3:v180 && sudo docker pull $t && sudo nvidia-docker run -it -
t=ultralytics/yolov3:v183 && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name 181 --cfg yolov3s9a-640.cfg
t=ultralytics/yolov3:v183 && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name 182 --cfg yolov3s9a-320-640.cfg
t=ultralytics/yolov3:v183 && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name 183 --cfg yolov3s15a-640.cfg
t=ultralytics/yolov3:v183 && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name 184 --cfg yolov3s15a-320-640.cfg
t=ultralytics/yolov3:v185 && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name 185
t=ultralytics/yolov3:v186 && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name 186
n=187 && t=ultralytics/yolov3:v$n && sudo docker pull $t && sudo nvidia-docker run -it --ipc=host -v "$(pwd)"/coco:/usr/src/coco $t python3 train.py --data coco2014.data --img-size 640 --epochs 10 --batch-size 22 --accumulate 3 --weights '' --arc defaultpw --pre --multi --bucket yolov4 --name $n

4
utils/utils.py
View File

@ -778,7 +778,7 @@ def kmean_anchors(path='../coco/train2017.txt', n=9, img_size=(320, 640)):
else:
# Kmeans calculation
from scipy.cluster.vq import kmeans
print('Running kmeans on %g points...' % len(wh))
print('Running kmeans for %g anchors on %g points...' % (n, len(wh)))
s = wh.std(0) # sigmas for whitening
k, dist = kmeans(wh / s, n, iter=20) # points, mean distance
k *= s
@ -800,7 +800,7 @@ def kmean_anchors(path='../coco/train2017.txt', n=9, img_size=(320, 640)):
fg = fitness(thr, wh, kg)
if fg > f:
f, k = fg, kg.copy()
print(fg, list(k.round().reshape(-1)))
print_results(thr, wh, k)
k = print_results(thr, wh, k)
return k