Gourav Roy 619ea0944e Avoid Memory Leak in Rollout worker ...

ISSUE: When we restore checkpoints, we create new nodes in the
Tensorflow graph. This happens when we assign new value (op node) to
RefVariable in GlobalVariableSaver. With every restore the size of TF
graph increases as new nodes are created and old unused nodes are not
removed from the graph. This causes the memory leak in
restore_checkpoint codepath.

FIX: We use TF placeholder to update the variables which avoids the
memory leak.

2019-01-02 23:09:09 -08:00

..

__init__.py

pre-release 0.10.0

2018-08-13 17:11:34 +03:00

test_basic_rl_graph_manager.py

Avoid Memory Leak in Rollout worker

2019-01-02 23:09:09 -08:00

test_graph_manager.py

introduce graph_manager.phase_context; make sure that calls to graph_manager.train automatically set training phase

2018-10-23 16:57:43 -04:00