Join TensorFlow at Google I/O, May 11-12 Register now


  • Description:

This data set contains roughly 44,000 examples of robot pushing motions, including one training set (train) and two test sets of previously seen (testseen) and unseen (testnovel) objects. This is the small 64x64 version.

Split Examples
'test' 256
'train' 43,264
  • Feature structure:
    'action': Tensor(shape=(4,), dtype=tf.float32),
    'endeffector_pos': Tensor(shape=(3,), dtype=tf.float32),
    'image_aux1': Image(shape=(64, 64, 3), dtype=tf.uint8),
    'image_main': Image(shape=(64, 64, 3), dtype=tf.uint8),
  • Feature documentation:
Feature Class Shape Dtype Description
action Tensor (4,) tf.float32
endeffector_pos Tensor (3,) tf.float32
image_aux1 Image (64, 64, 3) tf.uint8
image_main Image (64, 64, 3) tf.uint8
  • Citation:
  Author = {Frederik Ebert and Chelsea Finn and Alex X. Lee and Sergey Levine},
  Title = {Self-Supervised Visual Planning with Temporal Skip Connections},
  Year = {2017},
  Eprint = {arXiv:1710.05268},