search

luoyetx / deep-landmark

Predict facial landmarks with Deep CNNs powered by Caffe.
BSD 3-Clause "New" or "Revised" License
282 stars 156 forks source link

When I train the model, which is entered. / Bootstrap.sh always show sh: 1: caffe not found, what is the problem? How to deal with it? #35

Open 406baoqichao406 opened 7 years ago

406baoqichao406 commented 7 years ago

When I train the model, which is entered. / Bootstrap.sh always show sh: 1: caffe not found, what is the problem? How to deal with it?

liu09114 commented 7 years ago

Install caffe, and the command should be in py file should be "./build/tools/caffe train ...” in the caffe's root dir.

406baoqichao406 commented 7 years ago

You meant to put the deep landmark folder in "./build/tools/caffe train ..." and enter this command in the "./build/tools/caffe train" folder, right?

406baoqichao406 commented 7 years ago

I have compiled caffe, caffe environment variables have been on the computer, py file path has been changed over. Do py files have caffe related to the path?

huinsysu commented 6 years ago

Have you solved this problem? Please tell me the solution.Thanks!

huinsysu commented 6 years ago

I know how to solve it! Just change some codes in train/level.py. Replace "cmd = 'caffe train --solver prototxt/{0}_{1}solver.prototxt'.format(level, t)" with "cmd = '~/caffe/build/tools/caffe train --solver prototxt/{0}{1}solver.prototxt'.format(level, t)". And do the similar change to code "cmds.append('caffe train --solver prototxt/{0}{1}_solver.prototxt'.format(level, t))"