new model added

app.py

@@ -19,7 +19,7 @@ device = torch.device("cpu")
 model.to(device)
 
 transform = v2.Compose([
-        v2.Resize((96, 96)),
+        v2.Resize((32, 32)),
         v2.ToTensor(), 
         v2.Normalize(mean=[0.5, 0.5, 0.5], std=[0.5, 0.5, 0.5]),
     ])
@@ -53,21 +53,27 @@ def visualize_single_image(image_path):
 
     # MAE reconstruction pasted with visible patches
     im_paste = img * (1 - mask) + predicted_img * mask
+
+    # resize the image to 96 x 96
+    img = v2.functional.resize(img[0], (96, 96))
+    im_masked = v2.functional.resize(im_masked[0], (96, 96))
+    predicted_img = v2.functional.resize(predicted_img[0], (96, 96))
+    im_paste = v2.functional.resize(im_paste[0], (96, 96))
     
-   # make the plt figure larger
+    # make the plt figure larger
     plt.figure(figsize=(18, 8))
 
     plt.subplot(1, 4, 1)
-    show_image(img[0], "original")
+    show_image(img, "original")
 
     plt.subplot(1, 4, 2)
-    show_image(im_masked[0], "masked")
+    show_image(im_masked, "masked")
 
     plt.subplot(1, 4, 3)
-    show_image(predicted_img[0], "reconstruction")
+    show_image(predicted_img, "reconstruction")
 
     plt.subplot(1, 4, 4)
-    show_image(im_paste[0], "reconstruction + visible")
+    show_image(im_paste, "reconstruction + visible")
 
     plt.tight_layout()
 

vit-t-mae-pretrain.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0abe9d41d86e1dab46b8ae0139e6bfd5c3e909c41ccb7b7d7e77f8558e4f6db1
-size 29121704
+oid sha256:852a6a0806c42a8c725b0de82cd0e7b59d7d79ad21f8e012bc599eedcce15375
+size 28972154