diff --git a/.gitignore b/.gitignore
index 0d20b64..a86cb60 100644
--- a/.gitignore
+++ b/.gitignore
@@ -1 +1,7 @@
 *.pyc
+
+# checkpoint directories
+checkpoint/*
+
+# data
+data/*
\ No newline at end of file
diff --git a/geort/dataset.py b/geort/dataset.py
index 2166387..25ea4af 100644
--- a/geort/dataset.py
+++ b/geort/dataset.py
@@ -47,7 +47,7 @@ def __init__(self, qpos_keypoint_file, keypoint_names):
         self.qpos = np_array["qpos"]
         self.keypoints = np_array["keypoint"].item()
         self.keypoint_names = keypoint_names
-        print("Keypoint Names", self.keypoint_names)
+        # print("Keypoint Names", self.keypoint_names)  # Commented out to reduce verbose output
         self.n = len(self.qpos)
         return
     
diff --git a/geort/export.py b/geort/export.py
index 6b016c0..004374b 100644
--- a/geort/export.py
+++ b/geort/export.py
@@ -7,6 +7,7 @@
 import torch
 import os 
 from pathlib import Path
+import json
 from geort.formatter import HandFormatter
 from geort.model import IKModel
 from geort.utils.path import to_package_root, get_checkpoint_root
@@ -21,7 +22,7 @@ def __init__(self, model_path, config_path):
         config = load_json(config_path)
         keypoint_info = parse_config_keypoint_info(config)
         joint_lower_limit, joint_upper_limit = parse_config_joint_limit(config)
-        print(keypoint_info["joint"])
+        # print(keypoint_info["joint"])  # Commented out to reduce verbose output
         self.human_ids = keypoint_info["human_id"]
         self.model = IKModel(keypoint_joints=keypoint_info["joint"]).cuda()
         self.model.load_state_dict(torch.load(model_path))
@@ -36,9 +37,14 @@ def forward(self, keypoints):
         return joint_raw[0]
 
 
-def load_model(tag='', epoch=0):
+def load_model(tag='', epoch=0, use_best=True):
     '''
-        Loading API.
+        Loading API with best model preference.
+        
+        Args:
+            tag: checkpoint tag to search for
+            epoch: specific epoch to load (0 means latest)
+            use_best: if True, prefer best.pth over last.pth when epoch=0
     '''
     checkpoint_root = get_checkpoint_root()
     all_checkpoints = os.listdir(checkpoint_root)
@@ -50,10 +56,21 @@ def load_model(tag='', epoch=0):
             break 
 
     checkpoint_root = Path(checkpoint_root) / checkpoint_name
+    
     if epoch > 0:
         model_path = checkpoint_root / f"epoch_{epoch}.pth"
     else:
-        model_path = checkpoint_root / f"last.pth"
+        # Try to load best model first if use_best is True
+        if use_best and (checkpoint_root / "best.pth").exists():
+            model_path = checkpoint_root / "best.pth"
+            # Load training history to show best model info
+            history_path = checkpoint_root / "training_history.json"
+            if history_path.exists():
+                with open(history_path, 'r') as f:
+                    history = json.load(f)
+                    print(f"Loading best model from epoch {history['best_epoch']} with validation loss: {history['best_val_loss']:.4f}")
+        else:
+            model_path = checkpoint_root / "last.pth"
     
     config_path = checkpoint_root / "config.json"
     return GeoRTRetargetingModel(model_path=model_path, config_path=config_path)