Summary: 45 instances, 41 unique

Text	Count
# TODO: Not sure which reset to use here.	1
# TODO: vectorize this -- probably could be made faster	1
# TODO diagnostics per goal	1
self._training_environment.set_goal(self._goal_index) # TODO: implement in env	1
self._training_environment.env.set_goal(self._goal_index) # TODO: implement in env	1
# TODO: Why is the image being converted to float32 twice? Once in the	1
ImageId: <TODO:AMI>	2
# TODO: Hack	1
# TODO: Do this anti-aliasing in Mujoco render instead	1
# TODO: this currently only works for the rod	1
# TODO: figure out correct calibrations for these	1
# TODO: This is just defaulting to training all phases right now	1
# TODO: generalize this to diffenent padding types (only works for	1
# TODO: Check the hand is reset properly	1
ValueError: if TODO happens	1
# TODO: Fix this; redundant, since I'm going to reflatten it later	1
# TODO: Change this to phases	1
# TODO: Removed this due to releasing the object (move releasing for	1
# TODO: this logic of replacing the sawyer x euler velocity lim is sketchy	1
#TODO Avi maybe add temperature here	1
horizon = 100  # TODO: fix this	1
# TODO: How to handle asymmetric ranges where one side should be allowed	1
# TODO: Be able to specify convnet params, do this with `convnet.py`	1
# TODO: Is this correct logic? This would basically ignore the policy	1
# TODO: Clean up these two updates, lots of redundancy	1
# TODO: Create an abstract interface for preprocessors	1
# TODO: interleave videos from different policies	2
# TODO: Need to change this to bias things	1
# TODO: Need to fix the OptiTrack axes?	1
# TODO: Move this into object.py	1
new_goal_index = self._training_environment.env.task_graph() # TODO: Change this to call (check current phase)	1
# TODO: remnants from rllab -> gym conversion	3
## TODO: check if screw has moved within path	1
# TODO: Move these into base.py	1
# TODO: Implement this with `feedforward` network	1
# TODO: This needs testing	1
# TODO: Move this elsewhere	1
# TODO: test tf.nn.relu	1
# TODO: Figure out what in the environment is causing the Mujoco errors.	1
reposition_only=True,  # TODO: this flag is a little confusing	1
# TODO: Get this right	1