Using this sample with GPU support

[sa7ina]

Hi
And thanks for this repo.

I got it to work "OUT OF THE BOX" with tensorflow cpu 1.4.0 in the requirements.txt.
But it does not work with any other cpu version of tensorflow nor 1.4.0-gpu or any other gpu version.
Tried hard in the last couple days (even installed ubuntu more than 10 times breaking and starting all over again) but no success so far. only TF1.4.0 cpu.
Other examples i use works flawlessly with my GPU. GTX1080.

Any idea why and how to fix this excellent example code to work with gpu even newer version of tensorflow?

10X
--{-@

[sa7ina]

This is what i get when i use TF1.4.0-gpu.

`$ python eval.py
Traceback (most recent call last):
File "eval.py", line 20, in
import tensorflow as tf
File "/home/ei8ht/.local/lib/python2.7/site-packages/tensorflow/init.py", line 24, in
from tensorflow.python import *
File "/home/ei8ht/.local/lib/python2.7/site-packages/tensorflow/python/init.py", line 49, in
from tensorflow.python import pywrap_tensorflow
File "/home/ei8ht/.local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 72, in
raise ImportError(msg)
ImportError: Traceback (most recent call last):
File "/home/ei8ht/.local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in
from tensorflow.python.pywrap_tensorflow_internal import *
File "/home/ei8ht/.local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in
_pywrap_tensorflow_internal = swig_import_helper()
File "/home/ei8ht/.local/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
_mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
ImportError: libcublas.so.8.0: cannot open shared object file: No such file or directory

Failed to load the native TensorFlow runtime.`

[saket424]

@sa7ina
This error is because you done have CUDA 8.0 libraries installed

@igor-panteleev
I tried using tensorflow-gpu 1.5 since it uses CUDA 9.0 libraries but I get the following error. Can you take a look?

(va3) anand@Alienware15:~/objd/devicehive-video-analysis$ python eval.py
Traceback (most recent call last):
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_flag.py", line 166, in _parse
return self.parser.parse(argument)
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_argument_parser.py", line 114, in parse
type(argument)))
TypeError: flag value must be a string, found "<class 'int'>"

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "eval.py", line 123, in
tf.flags.DEFINE_string('video', 0, 'Path to the video file.')
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/tensorflow/python/platform/flags.py", line 58, in wrapper
return original_function(*args, **kwargs)
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_defines.py", line 241, in DEFINE_string
DEFINE(parser, name, default, help, flag_values, serializer, **args)
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_defines.py", line 81, in DEFINE
DEFINE_flag(_flag.Flag(parser, serializer, name, default, help, **args),
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_flag.py", line 107, in init
self._set_default(default)
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_flag.py", line 196, in _set_default
self.default = self._parse(value)
File "/home/anand/.virtualenvs/va3/lib/python3.6/site-packages/absl/flags/_flag.py", line 169, in _parse
'flag --%s=%s: %s' % (self.name, argument, e))
absl.flags._exceptions.IllegalFlagValueError: flag --video=0: flag value must be a string, found "<class 'int'>"

[MartinWeniger]

@saket424: i had the same Problem, i made it when i changed
tf.flags.DEFINE_string('video', 0, 'Path to the video file.') (the 0 is here an int)
to
tf.flags.DEFINE_string('video', '0', 'Path to the video file.') (the 0 is here an string)

And for Webcam mode i changed
cam = cv2.VideoCapture(Video)
to
cam = cv2.VideoCapture(0)

[MartinWeniger]

it runs for me under Windows 10 (64bit), with Tensorflow 1.8 gpu and python 3.5.5 with constant 8 FPS
i hopw it helps u

[saket424]

@MartinWeniger,
Thanks for this tips. With those two changes to eval.py that you outlined, I got it working using tensorflow-gpu 1.8 and I am getting 15fps on a nvidia geforce GTX 1080ti