[2023.05.29] commit-2

DarriusL · DarriusL · commit cbcab7a811cd · 2023-05-29T21:26:37.000+08:00
`Fixed the problem that each sampling of PER in the system is the first experience.
diff --git a/agent/algorithm/dqn.py b/agent/algorithm/dqn.py
@@ -92,9 +92,9 @@ def init_net(self, net_cfg, optim_cfg, lr_schedule_cfg, in_dim, out_dim, max_epo
         '''
         super().init_net(net_cfg, optim_cfg, lr_schedule_cfg, in_dim, out_dim, max_epoch);
         self.q_target_net = get_net(net_cfg, in_dim, out_dim).to(glb_var.get_value('device'));
-        self.net_updater.set_net(self.q_net, self.q_target_net);
         #Initialize q_target_net with q_net
         self.net_updater.net_param_copy(self.q_net, self.q_target_net);
+        self.net_updater.set_net(self.q_net, self.q_target_net);
         self.q_eval_net = self.q_target_net;
 
     def update(self):
diff --git a/agent/memory/offpolicy.py b/agent/memory/offpolicy.py
@@ -293,7 +293,7 @@ def get_stock(self):
     def sample(self):
         '''Sample batch'''
         if self.is_training:
-            if len(self.exps_latest) == 0:
+            if len(self.exps_latest) != 0:
                 #Add the experience that has not been added to the tree first.
                 idxs, priorities = tuple(zip(*self.exps_latest));
                 self.exps_latest.clear();
diff --git a/config/dqn/doubledqn_cartpole_per.json b/config/dqn/doubledqn_cartpole_per.json
@@ -6,14 +6,14 @@
                 "name":"linear",
                 "var_start":5.0,
                 "var_end":0.05,
-                "star_epoch":1000,
+                "star_epoch":100,
                 "end_epoch":10000
             },
             "gamma": 0.99,
             "net_updte_cfg":{
                 "name":"polyak",
                 "beta":0.5,
-                "update_step":10
+                "update_step":50
             }
         },
         "net_cfg":{
@@ -23,7 +23,7 @@
         },
         "optimizer_cfg":{
             "name":"adam",
-            "lr":2e-3,
+            "lr":1e-2,
             "weight_decay": 1e-08,
             "betas": [
                 0.9,
