Update README.md

Add some comments on pooling layer initialization according to https://github.com/huggingface/transformers/issues/25377#issuecomment-2164986543 🤗

cc

@amyeroberts



@nielsr

README.md

@@ -41,7 +41,10 @@ url = 'http://images.cocodataset.org/val2017/000000039769.jpg'
 image = Image.open(requests.get(url, stream=True).raw)
 
 processor = ViTImageProcessor.from_pretrained('facebook/dino-vits16')
-model = ViTModel.from_pretrained('facebook/dino-vits16')
+model = ViTModel.from_pretrained(
+    'facebook/dino-vits16',
+    add_pooling_layer=True  # Whether to initialize the ViTPooler module. Set it to 'False' to load only the backbone. 🤗
+)
 
 inputs = processor(images=image, return_tensors="pt")
 outputs = model(**inputs)