gryf/coach
1
0
Fork 0
mirror of https://github.com/gryf/coach.git synced 2025-12-17 19:20:19 +01:00
Code Issues Projects Releases Wiki Activity

appending csv's from logger instead of rewriting them

Browse Source
This commit is contained in:
Itai Caspi
2018-02-12 01:33:43 +02:00
committed by Itai Caspi
parent 569ca39ce6
commit ba96e585d2
6 changed files with 33 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
agents/n_step_q_agent.py
View File

@@ -28,8 +28,10 @@ class NStepQAgent(ValueOptimizationAgent, PolicyOptimizationAgent):
self.last_gradient_update_step_idx = 0
self.q_values = Signal('Q Values')
self.unclipped_grads = Signal('Grads (unclipped)')
self.value_loss = Signal('Value Loss')
self.signals.append(self.q_values)
self.signals.append(self.unclipped_grads)
self.signals.append(self.value_loss)
def learn_from_batch(self, batch):
# batch contains a list of episodes to learn from
@@ -69,7 +71,7 @@ class NStepQAgent(ValueOptimizationAgent, PolicyOptimizationAgent):
# logging
total_loss, losses, unclipped_grads = result[:3]
self.unclipped_grads.add_sample(unclipped_grads)
logger.create_signal_value('Value Loss', losses[0])
self.value_loss.add_sample(losses[0])
return total_loss