0

I am trying to use tf slim library for building model on top of inception-v3 model

The tf slim library uses a version 'import tensorflow.contrib.slim as slim'

wheras the inception model uses 'from inception.slim import slim'

both the versions have some common and some different set of files , like layers.py is not present in inception.slim etc

  1. What is the significance of having two versions of the library ?
  2. Is there a difference in application of both the libraries
  3. The inception.slim contains files with deprecated functions like tf.op_scope and tf.variable_op_scope ?
Ashish Kumar
  • 23
  • 6

1 Answers1

1
  1. We are moving layers to the core library.
  2. There may be subtle differences, the core library is always the recommended one.
  3. Yes, while slim/tf.learn etc are being assimilated, they will be in an odd shape until then. (But they should continue to work)
drpng
  • 1,637
  • 13
  • 14
  • could you also explain the difference/similarity between tflearn and tf.slim ?? both seem to be high level api's for tf .Thanks – Ashish Kumar Nov 15 '16 at 08:05
  • They used to be different ways of interacting with tensorflow. Technically, slim is a part of tflearn. We are cherry-picking them into the core, and ultimately all functionality at different levels of granularity should be available with the core. – drpng Nov 15 '16 at 13:00