diff --git a/datasets/sky_datasets.py b/datasets/sky_datasets.py
index b5d6f22..6f135a9 100644
--- a/datasets/sky_datasets.py
+++ b/datasets/sky_datasets.py
@@ -54,7 +54,7 @@ def load_video_frames(self, dataroot):
             root = meta[0]
             try:
                 frames = sorted(meta[2], key=lambda item: int(item.split('.')[0].split('_')[-1]))
-            except:
+            except ValueError:
                 print(meta[0]) # root
                 print(meta[2]) # files
             frames = [os.path.join(root, item) for item in frames if is_image_file(item)]
diff --git a/datasets/sky_image_datasets.py b/datasets/sky_image_datasets.py
index 8ac3c70..857a64b 100644
--- a/datasets/sky_image_datasets.py
+++ b/datasets/sky_image_datasets.py
@@ -78,7 +78,7 @@ def load_video_frames(self, dataroot):
             root = meta[0]
             try:
                 frames = sorted(meta[2], key=lambda item: int(item.split('.')[0].split('_')[-1]))
-            except:
+            except ValueError:
                 print(meta[0]) # root
                 print(meta[2]) # files
             frames = [os.path.join(root, item) for item in frames if is_image_file(item)]
diff --git a/datasets/taichi_datasets.py b/datasets/taichi_datasets.py
index 9d64b43..4648869 100644
--- a/datasets/taichi_datasets.py
+++ b/datasets/taichi_datasets.py
@@ -57,7 +57,7 @@ def load_video_frames(self, dataroot):
             root = meta[0]
             try:
                 frames = sorted(meta[2], key=lambda item: int(item.split('.')[0].split('_')[-1]))
-            except:
+            except ValueError:
                 print(meta[0], meta[2])
             frames = [os.path.join(root, item) for item in frames if is_image_file(item)]
             # if len(frames) > max(0, self.sequence_length * self.sample_every_n_frames):
diff --git a/datasets/taichi_image_datasets.py b/datasets/taichi_image_datasets.py
index d657dcb..18544ee 100644
--- a/datasets/taichi_image_datasets.py
+++ b/datasets/taichi_image_datasets.py
@@ -84,7 +84,7 @@ def load_video_frames(self, dataroot):
             root = meta[0]
             try:
                 frames = sorted(meta[2], key=lambda item: int(item.split('.')[0].split('_')[-1]))
-            except:
+            except ValueError:
                 print(meta[0], meta[2])
             frames = [os.path.join(root, item) for item in frames if is_image_file(item)]
             # if len(frames) > max(0, self.sequence_length * self.sample_every_n_frames):