diff --git a/celeste/main.py b/celeste/main.py
index 5b27be5..4f6e5f8 100644
--- a/celeste/main.py
+++ b/celeste/main.py
@@ -9,8 +9,17 @@ import torch
 from celeste import Celeste
 
 
-run_data_path = Path("out")
-run_data_path.mkdir(parents = True, exist_ok = True)
+# Where to read/write model data.
+model_data_root = Path("model_data")
+
+model_save_path		= model_data_root / "model.torch"
+model_archive_dir	= model_data_root / "model_archive"
+model_train_log		= model_data_root / "train_log"
+screenshot_dir		= model_data_root / "screenshots"
+model_data_root.mkdir(parents = True, exist_ok = True)
+model_archive_dir.mkdir(parents = True, exist_ok = True)
+screenshot_dir.mkdir(parents = True, exist_ok = True)
+
 
 compute_device = torch.device(
 	"cuda" if torch.cuda.is_available() else "cpu"
@@ -307,9 +316,9 @@ def optimize_model():
 episode_number = 0
 
 
-if (run_data_path/"checkpoint.torch").is_file():
+if model_save_path.is_file():
 	# Load model if one exists
-	checkpoint = torch.load((run_data_path/"checkpoint.torch"))
+	checkpoint = torch.load(model_save_path)
 	policy_net.load_state_dict(checkpoint["policy_state_dict"])
 	target_net.load_state_dict(checkpoint["target_state_dict"])
 	optimizer.load_state_dict(checkpoint["optimizer_state_dict"])
@@ -437,7 +446,7 @@ def on_state_after(celeste, before_out):
 	# a terminal state.
 	if (next_state.deaths != 0):
 		s = celeste.state
-		with open(run_data_path / "train.log", "a") as f:
+		with model_train_log.open("a") as f:
 			f.write(json.dumps({
 				"checkpoints": s.next_point,
 				"state_count": s.state_count
@@ -452,13 +461,13 @@ def on_state_after(celeste, before_out):
 			"memory": memory,
 			"episode_number": episode_number,
 			"steps_done": steps_done
-		}, run_data_path / "checkpoint.torch")
+		}, model_save_path)
 
 
 		# Clean up screenshots
 		shots = Path("/home/mark/Desktop").glob("hackcel_*.png")
 
-		target = run_data_path / Path(f"screenshots/{episode_number}")
+		target = screenshot_dir / Path(f"{episode_number}")
 		target.mkdir(parents = True)
 
 		for s in shots:
@@ -466,8 +475,6 @@ def on_state_after(celeste, before_out):
 
 		# Save a prediction graph
 		if episode_number % image_interval == 0:
-			p =	run_data_path / Path("model_images")
-			p.mkdir(parents = True, exist_ok = True)
 			torch.save({
 				"policy_state_dict": policy_net.state_dict(),
 				"target_state_dict": target_net.state_dict(),
@@ -475,10 +482,10 @@ def on_state_after(celeste, before_out):
 				"memory": memory,
 				"episode_number": episode_number,
 				"steps_done": steps_done
-			}, p / f"{episode_number}.torch")
+			}, model_archive_dir / f"{episode_number}.torch")
 
 
-		print("State over, resetting")
+		print("Game over. Resetting.")
 		episode_number += 1
 		celeste.reset()