Fix some bugs

src/qpd/compression/memory.py

@@ -118,7 +118,7 @@ class Memory:
         self.max_size += amount
         self.did_overflow = False
 
-        if self.verifyConsistency:
+        if self.check_consistency:
             self.verifyConsistency("Grow")
         return len(self)
 

src/qpd/compression/policy_distillation.py

@@ -1,4 +1,5 @@
 # import wandb
+import copy
 import random
 import os
 import gc
@@ -18,7 +19,7 @@ from torch.distributions import Categorical
 import torch
 import torch.nn as nn
 import torch.nn.functional as F
-from typing import Dict, Optional, Union
+from typing import Dict, Optional, Tuple, Union
 
 from .memory import Memory
 
@@ -76,6 +77,7 @@ class PolicyDistillation:
             if (epoch and epoch % self.config.evaluator_config.student_test_frequency == 0) or self.config.evaluator_config.student_test_frequency == 1:
                 best, is_best = self.test(epoch, best)
                 if is_best:
+                    self.best_student = copy.deepcopy(self.student)
                     self.save_student("best")
 
             if (epoch and epoch % self.check_point_frequency == 0) or self.check_point_frequency == 1:
@@ -236,7 +238,7 @@ class PolicyDistillation:
         # if self.webdav_manager:
         #     self.webdav_manager.upload_model(student_save_path, epoch, not self.config["testing"]["local_checkpoints"])
         
-    def test(self, epoch: int, best: Optional[float]=None) -> Optional[float]:
+    def test(self, epoch: int, best: Optional[float]=None) -> Tuple[Optional[float], bool]:
         is_best = False
         print()
         print("= Testing =========================================================")

src/qpd/utils/evaluator.py

@@ -162,7 +162,7 @@ class Evaluator:
                         action, value, action_std = teacher(state)
 
                     if not worker_finished:
-                        if worker_memory.will_overflow:
+                        if worker_memory.will_overflow():
                             print(f"Growing memory to size {worker_memory.max_size + 5000}!")
                             worker_memory.grow(5000)
 
@@ -191,7 +191,7 @@ class Evaluator:
                         total_steps.append(worker_steps)
                         worker_steps = 0
 
-                        if (steps and effective_frames >= steps) or (episodes and len(total_rewards)):
+                        if (steps and effective_frames >= steps) or (episodes and len(total_rewards) >= episodes):
                             worker_finished = True
 
         if collect:
@@ -201,7 +201,7 @@ class Evaluator:
             end = time.time()
             print("Average and effective framerate:", round(frames / (end-start), 3), round(effective_frames / (end-start), 3))
 
-        return {"Steps": numpy.mean(total_steps) if len(total_steps) else 0, "Reward": numpy.mean(total_rewards) if len(total_rewards) else 0, "Memory": worker_memory if collect else []}
+        return {"Steps": numpy.mean(total_steps) if len(total_steps) else 0, "Reward": numpy.mean(total_rewards) if len(total_rewards) else 0, "Memory": worker_memory if collect else [], "EpisodeCount": len(total_rewards)}
 
     @abstractmethod
     def test_vec(config: Config,