frame

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Alert: Beginning Tuesday, June 25th, we will be freezing this site and migrating the content and forums to our new home at https://forums.intel.com. Check it out now!

When using the TF data api and iterators, how do you name the input?

Here is how I input data into my network:

iterator = TFRecordParser.input_fn(batch_size, record_path)
sess.run(iterator.initializer)
image, ground_truth = iterator.get_next()
prediction = network(image)

So "image" is the data going into my network, and "prediction" is the output. How do I name these so that I can compile with mvNCCompile?

Comments

  • 1 Comment sorted by Votes Date Added
  • Nevermind, I figured it out. Basically, when you go to make your inference only python file, you have to delete the iterator code anyways because you don't need it. So just replace

    image, ground_truth = iterator.get_next()

    with

    image = tf.placeholder(tf.float32, [None, 480, 640, 3], name="input")

    even if you didn't use the feed_dict pattern.

Sign In or Register to comment.