diff --git a/deepem/train/sample.py b/deepem/train/sample.py
new file mode 100644
index 0000000..2ee8bf4
--- /dev/null
+++ b/deepem/train/sample.py
@@ -0,0 +1,67 @@
+import os
+import sys
+import time
+
+import numpy as np
+import torch
+import samwise
+import napari
+
+from deepem.train.option import Options
+from deepem.train.utils import *
+
+
+def sample(opt):
+
+    if opt.batch_size != 1 or opt.num_workers != 1:
+        warnings.warn("Setting batch size and # workers to 1 for now")
+        opt.batch_size = 1
+        opt.num_workers = 1
+
+    # Data loaders
+    train_loader, val_loader = load_data(opt)
+
+    # Sample loop
+    t0 = time.time()
+
+    viewer = napari.Viewer()
+
+    @viewer.bind_key("q")
+    def quit(viewer):
+        sys.exit()
+
+    def convert(t: torch.Tensor) -> np.ndarray:
+        return t.cpu().detach().numpy().astype(np.uint8)
+
+    print("Press ENTER to continue, Q to quit")
+    for i in range(opt.chkpt_num, opt.max_iter):
+
+        with torch.no_grad():
+            # Load training samples.
+            sample = train_loader()
+
+            # Elapsed time
+            elapsed = time.time() - t0
+            print(f"Sample generated in {elapsed:.3f}s") 
+
+            # Viewer
+            viewer.add_image(convert(sample["input"] * 255), name="image")
+            for k in opt.out_spec:
+                viewer.add_labels(convert(sample[k]), name=k, opacity=0.5)
+                viewer.add_image(
+                    convert(sample[f"{k}_mask"] * 255), name=f"{k}_mask", opacity=0.2
+                )
+
+            # Wait for user input (enter to continue, q to quit)
+            input()
+
+            # Reset timer and viewer.
+            for layer in range(len(viewer.layers)):
+                viewer.layers.pop()
+            t0 = time.time()
+
+
+if __name__ == "__main__":
+    # Options
+    opt = Options().parse()
+    sample(opt)
diff --git a/requirements.txt b/requirements.txt
index 4b575a0..e4d292c 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -5,3 +5,4 @@ task-queue
 tensorflow
 tensorboard
 tensorboardX
+napari