train my own data

[sankim90]

Hi

• My class CLASSES = ('background', 'person', 'bicycle', 'car', 'motorcycle', 'bus', 'truck', 'traffic light', 'stop sign', 'dog'

• run gen_model.sh 10

• train.sh is

  !/bin/sh

  if ! test -f example/MobileNetSSD_train.prototxt ;then echo "error: example/MobileNetSSD_train.prototxt does not exist." echo "please use the gen_model.sh to generate your own model." exit 1 fi mkdir -p snapshot /home/sankim/caffe-ssd/build/tools/caffe train -solver="solver_train.prototxt" \ -gpu 0

• run demo.py net_file= 'MobileNetSSD_deploy.prototxt'
  caffe_model='mobilenet_iter_30000.caffemodel'

The train is over but the model performs incorrect detection

Do I have to add -weight MobileNetSSD_deploy.caffemodel to train.sh?

[soufianesabiri]

Were you able to train your model? If so please show me the steps... Thanks!

[mibrahimy]

@sankim90 were you able to know what was the issue?

[sankim90]

@soufianesabiri @ibrahimsherdil Nope, I did not use the MobileNet SSD because I thought the model was unreliable.