dd

app.py

@@ -7,34 +7,34 @@ import numpy as np
 
 
 def greet(url):
-
+    # load Mask2Former fine-tuned on Cityscapes semantic segmentation
     processor = AutoImageProcessor.from_pretrained("facebook/mask2former-swin-large-cityscapes-semantic")
     model = Mask2FormerForUniversalSegmentation.from_pretrained("facebook/mask2former-swin-large-cityscapes-semantic")
 
     image = Image.open(requests.get(url, stream=True).raw)
+
     inputs = processor(images=image, return_tensors="pt")
 
     with torch.no_grad():
         outputs = model(**inputs)
+
     # model predicts class_queries_logits of shape `(batch_size, num_queries)`
     # and masks_queries_logits of shape `(batch_size, num_queries, height, width)`
     class_queries_logits = outputs.class_queries_logits
     masks_queries_logits = outputs.masks_queries_logits
 
     # you can pass them to processor for postprocessing
-    predicted_semantic_map = processor.post_process_semantic_segmentation(outputs, target_sizes=[image.size])[0]
-    predicted_semantic_map = predicted_semantic_map.numpy()
-
-    return predicted_semantic_map
-
+    predicted_semantic_map = processor.post_process_semantic_segmentation(outputs, target_sizes=[image.size[::-1]])[0]
+    # we refer to the demo notebooks for visualization (see "Resources" section in the Mask2Former docs)
 
-url = "http://images.cocodataset.org/val2017/000000039769.jpg"
+url = "http://www.apparelnews.co.kr/upfiles/manage/202302/5d5f694177b26fc86e5db623bf7ae4b7.jpg"
 #greet(url)
 
 iface = gr.Interface(
     fn=greet,
     inputs=gr.Image(value=url),
-    outputs="image"
+    outputs=gr.Image(),
+    live=True
 )
 
 iface.launch(debug = True)