adirik commited on
Commit
9e1896f
1 Parent(s): 67fc4e8

remove manual resizing, add example

Browse files
Files changed (3) hide show
  1. app.py +11 -5
  2. assets/.DS_Store +0 -0
  3. assets/butterflies.jpeg +0 -0
app.py CHANGED
@@ -21,9 +21,7 @@ def query_image(img, text_queries, score_threshold):
21
  text_queries = text_queries.split(",")
22
 
23
  target_sizes = torch.Tensor([img.shape[:2]])
24
- img_input = cv2.resize(img, (768, 768), interpolation = cv2.INTER_AREA)
25
-
26
- inputs = processor(text=text_queries, images=img_input, return_tensors="pt").to(device)
27
 
28
  with torch.no_grad():
29
  outputs = model(**inputs)
@@ -57,7 +55,11 @@ introduced in <a href="https://arxiv.org/abs/2205.06230">Simple Open-Vocabulary
57
  with Vision Transformers</a>.
58
  \n\nYou can use OWL-ViT to query images with text descriptions of any object.
59
  To use it, simply upload an image and enter comma separated text descriptions of objects you want to query the image for. You
60
- can also use the score threshold slider to set a threshold to filter out low probability predictions.
 
 
 
 
61
  \n\n<a href="https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/zeroshot_object_detection_with_owlvit.ipynb">Colab demo</a>
62
  """
63
  demo = gr.Interface(
@@ -66,6 +68,10 @@ demo = gr.Interface(
66
  outputs="image",
67
  title="Zero-Shot Object Detection with OWL-ViT",
68
  description=description,
69
- examples=[["assets/astronaut.png", "human face, rocket, flag, nasa badge", 0.11], ["assets/coffee.png", "coffee mug, spoon, plate", 0.1]],
 
 
 
 
70
  )
71
  demo.launch()
 
21
  text_queries = text_queries.split(",")
22
 
23
  target_sizes = torch.Tensor([img.shape[:2]])
24
+ inputs = processor(text=text_queries, images=img, return_tensors="pt").to(device)
 
 
25
 
26
  with torch.no_grad():
27
  outputs = model(**inputs)
 
55
  with Vision Transformers</a>.
56
  \n\nYou can use OWL-ViT to query images with text descriptions of any object.
57
  To use it, simply upload an image and enter comma separated text descriptions of objects you want to query the image for. You
58
+ can also use the score threshold slider to set a threshold to filter out low probability predictions.
59
+
60
+ \n\nOWL-ViT is trained on text templates,
61
+ hence you can get better predictions by querying the image with text templates used in training the original model: *"photo of a star-spangled banner"*,
62
+ *"image of a shoe"*. Refer to the <a href="https://arxiv.org/abs/2103.00020">CLIP</a> paper to see the full list of text templates used to augment the training data.
63
  \n\n<a href="https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/zeroshot_object_detection_with_owlvit.ipynb">Colab demo</a>
64
  """
65
  demo = gr.Interface(
 
68
  outputs="image",
69
  title="Zero-Shot Object Detection with OWL-ViT",
70
  description=description,
71
+ examples=[
72
+ ["assets/astronaut.png", "human face, rocket, star-spangled banner, nasa badge", 0.11],
73
+ ["assets/coffee.png", "coffee mug, spoon, plate", 0.1],
74
+ ["assets/butterflies.jpeg", "orange butterfly", 0.3],
75
+ ],
76
  )
77
  demo.launch()
assets/.DS_Store CHANGED
Binary files a/assets/.DS_Store and b/assets/.DS_Store differ
 
assets/butterflies.jpeg ADDED