policy linear lr

4 年前 · 2213a071
--- a/config/ppo_transfer/CrawlerStatic.yaml
+++ b/config/ppo_transfer/CrawlerStatic.yaml
      epsilon: 0.2
      lambd: 0.95
      num_epoch: 3
-      learning_rate_schedule: constant
+      learning_rate_schedule: linear
      model_schedule: linear
      encoder_layers: 2
      action_layers: 2
--- a/config/ppo_transfer/OldCrawlerStatic.yaml
+++ b/config/ppo_transfer/OldCrawlerStatic.yaml
      epsilon: 0.2
      lambd: 0.95
      num_epoch: 3
-      learning_rate_schedule: constant
+      learning_rate_schedule: linear
      model_schedule: linear
      encoder_layers: 2
      action_layers: 2