Vivien commited on
Commit
81c0cb7
β€’
1 Parent(s): 563e3ef

Switch to the large models

Browse files
Files changed (2) hide show
  1. README.md +1 -1
  2. app.py +7 -5
README.md CHANGED
@@ -1,5 +1,5 @@
1
  ---
2
- title: Search and Detect (CLIP/Owl-ViT)
3
  emoji: πŸ¦‰
4
  colorFrom: indigo
5
  colorTo: red
 
1
  ---
2
+ title: Search and Detect (CLIP/OWL-ViT)
3
  emoji: πŸ¦‰
4
  colorFrom: indigo
5
  colorTo: red
app.py CHANGED
@@ -12,7 +12,7 @@ from transformers import OwlViTProcessor, OwlViTForObjectDetection
12
  from transformers.image_utils import ImageFeatureExtractionMixin
13
  import tokenizers
14
 
15
- DEBUG = True
16
  if DEBUG:
17
  MODEL = "vit-base-patch32"
18
  OWL_MODEL = f"google/owlvit-base-patch32"
@@ -31,7 +31,7 @@ N_RESULTS = 6
31
 
32
  color = st.get_option("theme.primaryColor")
33
  if color is None:
34
- color = (255, 75, 75)
35
  else:
36
  color = tuple(int(color.lstrip("#")[i : i + 2], 16) for i in (0, 2, 4))
37
 
@@ -215,7 +215,7 @@ This demo illustrates how you can both retrieve images containing certain object
215
 
216
  **Tip 2**: change the score threshold below to adjust the sensitivity of the object detection.
217
 
218
- *Built with OpenAI's [CLIP](https://openai.com/blog/clip/) model and Google's [Owl-ViT](https://arxiv.org/abs/2205.06230) model, πŸ€— Hugging Face's [transformers library](https://huggingface.co/transformers/), [Streamlit](https://streamlit.io/), 25k images from [Unsplash](https://unsplash.com/) and 8k images from [The Movie Database (TMDB)](https://www.themoviedb.org/)*
219
 
220
  """
221
 
@@ -266,13 +266,15 @@ def main():
266
  )
267
 
268
  _, c, _ = st.columns((1, 3, 1))
269
- query = c.text_input("", value="clouds at sunset")
270
  corpus = st.radio("", ["Unsplash", "Movies"])
271
 
272
  if len(query) > 0:
273
  if "/" in query:
274
  queries = query.split("/")
275
- clip_query, owl_query = ("/").join(queries[:-1]), queries[-1]
 
 
276
  else:
277
  clip_query, owl_query = query, query
278
  retrieved = image_search(clip_query, corpus)
 
12
  from transformers.image_utils import ImageFeatureExtractionMixin
13
  import tokenizers
14
 
15
+ DEBUG = False
16
  if DEBUG:
17
  MODEL = "vit-base-patch32"
18
  OWL_MODEL = f"google/owlvit-base-patch32"
 
31
 
32
  color = st.get_option("theme.primaryColor")
33
  if color is None:
34
+ color = (255, 196, 35)
35
  else:
36
  color = tuple(int(color.lstrip("#")[i : i + 2], 16) for i in (0, 2, 4))
37
 
 
215
 
216
  **Tip 2**: change the score threshold below to adjust the sensitivity of the object detection.
217
 
218
+ *Built with OpenAI's [CLIP](https://openai.com/blog/clip/) model and Google's [OWL-ViT](https://arxiv.org/abs/2205.06230) model, πŸ€— Hugging Face's [transformers library](https://huggingface.co/transformers/), [Streamlit](https://streamlit.io/), 25k images from [Unsplash](https://unsplash.com/) and 8k images from [The Movie Database (TMDB)](https://www.themoviedb.org/)*
219
 
220
  """
221
 
 
266
  )
267
 
268
  _, c, _ = st.columns((1, 3, 1))
269
+ query = c.text_input("", value="koala")
270
  corpus = st.radio("", ["Unsplash", "Movies"])
271
 
272
  if len(query) > 0:
273
  if "/" in query:
274
  queries = query.split("/")
275
+ clip_query, owl_query = ("/").join(queries[:-1]).strip(), queries[
276
+ -1
277
+ ].strip()
278
  else:
279
  clip_query, owl_query = query, query
280
  retrieved = image_search(clip_query, corpus)