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

Batch RL (#238)

Browse Source
This commit is contained in:
Gal Leibovich
2019-03-19 18:07:09 +02:00
committed by GitHub
parent 4a8451ff02
commit e3c7e526c7
38 changed files with 1003 additions and 87 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
rl_coach/agents/n_step_q_agent.py
View File

@@ -123,6 +123,9 @@ class NStepQAgent(ValueOptimizationAgent, PolicyOptimizationAgent):
else:
assert True, 'The available values for targets_horizon are: 1-Step, N-Step'
# add Q value samples for logging
self.q_values.add_sample(state_value_head_targets)
# train
result = self.networks['main'].online_network.accumulate_gradients(batch.states(network_keys), [state_value_head_targets])