diff --git a/README.md b/README.md
index 02bb125..043e29c 100644
--- a/README.md
+++ b/README.md
@@ -17,5 +17,5 @@ pip3 install opencv-python
 
 ## Usage
 ```
-python3 inference.py --img img0.png img1.png
+python3 inference.py --img /path/to/image_0 /path/to/image_1
 ```
diff --git a/inference.py b/inference.py
index 9553f8a..e669fcb 100644
--- a/inference.py
+++ b/inference.py
@@ -1,16 +1,15 @@
 import cv2
 import torch
 import argparse
+from torch.nn import functional as F
 from model.RIFE import Model
 
 device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
 
 parser = argparse.ArgumentParser(description='Interpolation for a pair of images')
-parser.add_argument('--img', dest='img', nargs=2)
+parser.add_argument('--img', dest='img', nargs=2, required=True)
 args = parser.parse_args()
 
-assert(len(args.img) == 2)
-
 model = Model()
 model.load_model('./train_log')
 model.eval()
@@ -18,9 +17,12 @@ model.device()
 img0 = cv2.imread(args.img[0])
 img1 = cv2.imread(args.img[1])
 h, w, _ = img0.shape
+ph = h // 32 * 32
+pw = w // 32 * 32
+padding = (0, pw - w, 0, ph - h)
 img0 = torch.tensor(img0.transpose(2, 0, 1)).to(device) / 255.
 img1 = torch.tensor(img1.transpose(2, 0, 1)).to(device) / 255.
-imgs = torch.cat((img0, img1), 0).float()
+imgs = F.pad(torch.cat((img0, img1), 0).float(), padding)
 with torch.no_grad():
     res = model.inference(imgs.unsqueeze(0)) * 255
-cv2.imwrite('output.png', res[0].numpy().transpose(1, 2, 0))
+cv2.imwrite('output.png', res[0].numpy().transpose(1, 2, 0)[:h, :w])