diff --git a/map2map/data/fields.py b/map2map/data/fields.py
index 19d1ddb..e36dc99 100644
--- a/map2map/data/fields.py
+++ b/map2map/data/fields.py
@@ -135,20 +135,20 @@ class FieldDataset(Dataset):
             # first add full fields when num_fields > num_GPU
             for i in range(rank, self.nfile // world_size * world_size,
                            world_size):
-                self.samples.append(
+                self.samples.extend(list(
                     range(i * self.ncrop, (i + 1) * self.ncrop)
-                )
+                ))
 
             # then split the rest into fractions of fields
             # drop the last incomplete batch of samples
             frac_start = self.nfile // world_size * world_size * self.ncrop
             frac_samples = self.nfile % world_size * self.ncrop // world_size
-            self.samples.append(range(frac_start + rank * frac_samples,
-                                      frac_start + (rank + 1) * frac_samples))
-
-            self.samples = np.concatenate(self.samples)
+            self.samples.extend(list(
+                range(frac_start + rank * frac_samples,
+                      frac_start + (rank + 1) * frac_samples)
+            ))
         else:
-            self.samples = np.arange(self.nfile * self.ncrop)
+            self.samples = list(range(self.nfile * self.ncrop))
         self.nsample = len(self.samples)
 
         self.rank = rank