gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2025-12-17 11:10:20 +01:00
Code Issues Projects Releases Wiki Activity

Fixes for having NumpySharedRunningStats syncing on multi-node (#139)

Browse Source 
1. Having the standard checkpoint prefix in order for the data store to grab it, and sync it to S3.
2. Removing the reference to Redis so that it won't try to pickle that in.
3. Enable restoring a checkpoint into a single-worker run, which was saved by a single-node-multiple-worker run.
This commit is contained in:
Gal Leibovich
2018-11-23 16:11:47 +02:00
committed by GitHub
parent 87a7848b0a
commit a1c56edd98
12 changed files with 154 additions and 99 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
rl_coach/agents/nec_agent.py
View File

@@ -203,7 +203,7 @@ class NECAgent(ValueOptimizationAgent):
self.networks['main'].online_network.output_heads[0].DND.add(self.current_episode_state_embeddings,
actions, discounted_rewards)
def save_checkpoint(self, checkpoint_id):
super().save_checkpoint(checkpoint_id)
with open(os.path.join(self.ap.task_parameters.checkpoint_save_dir, str(checkpoint_id) + '.dnd'), 'wb') as f:
def save_checkpoint(self, checkpoint_prefix):
super().save_checkpoint(checkpoint_prefix)
with open(os.path.join(self.ap.task_parameters.checkpoint_save_dir, str(checkpoint_prefix) + '.dnd'), 'wb') as f:
pickle.dump(self.networks['main'].online_network.output_heads[0].DND, f, pickle.HIGHEST_PROTOCOL)