尝试使用张量流对我自己的图像进行分类

Question

I have just downloaded tensorflow and I'm trying to classify some images of my own through the command line.

What I've done until now is:

    $ sudo apt-get install python-pip python-dev

    # Ubuntu/Linux 64-bit, CPU only, Python 2.7
    $ export TF_BINARY_URL=https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.11.0rc2-cp27-none-linux_x86_64.whl

    sudo pip install --upgrade $TF_BINARY_URL

    #and finally, to install and run tensorflow classifier
    $python classify_image.py

which worked fine. However I cannot find how to supply my own images for classification. I'm aware of the --image_file filename.jpg however I cannot find where the images are supposed to be stored in order for this to work. I've tried giving the absolute path,

     $python classify_image.py --image_file /home/usr/Desktop/ball.jpg

but I get a "file does not exist" warning.

Full error: http://imgur.com/wcn2QlI Interesting thing is that now that I restarted the virtual machine it re-downloaded the model for some reason.

Is there something more to this, do I have to change the source code in some way?

Thanks in advance

Answer 1

The scripts, including examples, are all stored in your pip installation path.

Note that you have to pip install before: sudo pip install --upgrade $TF_BINARY_URL .

For instance, as above, it is probably here: ./usr/local/lib/python2.7/dist-packages/tensorflow/models/image/imagenet/classify_image.py .