Inception v3 in TensorFlow

Note

You can follow along with the code in the Jupyter notebook ch-12c_InceptionV3_TensorFlow.

TensorFlow's Inception v3 is trained on 1,001 labels instead of 1,000. Also, the images used for training are pre-processed differently. We showed the preprocessing code in previous sections. Let us dive directly into restoring the Inception v3 model using TensorFlow.

Let us download the checkpoint file for the Inception v3:

# load the inception V3 model
model_name='inception_v3'
model_url='http://download.tensorflow.org/models/'
model_files=['inception_v3_2016_08_28.tar.gz']
model_home=os.path.join(models_root,model_name) 

dsu.download_dataset(source_url=model_url,
    source_files=model_files,
    dest_dir = model_home,
    force=False,
    extract=True)

Define the common imports for inception module and variables:

### define common imports and variables
from tensorflow.contrib.slim.nets import inception
image_height=inception.inception_v3.default_image_size
image_width...