Switch to the large models

README.md

@@ -1,5 +1,5 @@
 ---
-title: Search and Detect (CLIP/Owl-ViT)
+title: Search and Detect (CLIP/OWL-ViT)
 emoji: 🦉
 colorFrom: indigo
 colorTo: red

app.py

@@ -12,7 +12,7 @@ from transformers import OwlViTProcessor, OwlViTForObjectDetection
 from transformers.image_utils import ImageFeatureExtractionMixin
 import tokenizers
 
-DEBUG = True
+DEBUG = False
 if DEBUG:
     MODEL = "vit-base-patch32"
     OWL_MODEL = f"google/owlvit-base-patch32"
@@ -31,7 +31,7 @@ N_RESULTS = 6
 
 color = st.get_option("theme.primaryColor")
 if color is None:
-    color = (255, 75, 75)
+    color = (255, 196, 35)
 else:
     color = tuple(int(color.lstrip("#")[i : i + 2], 16) for i in (0, 2, 4))
 
@@ -215,7 +215,7 @@ This demo illustrates how you can both retrieve images containing certain object
 
 **Tip 2**: change the score threshold below to adjust the sensitivity of the object detection.
 
-*Built with OpenAI's [CLIP](https://openai.com/blog/clip/) model and Google's [Owl-ViT](https://arxiv.org/abs/2205.06230) model, 🤗 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/)*
+*Built with OpenAI's [CLIP](https://openai.com/blog/clip/) model and Google's [OWL-ViT](https://arxiv.org/abs/2205.06230) model, 🤗 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/)*
 
 """
 
@@ -266,13 +266,15 @@ def main():
     )
 
     _, c, _ = st.columns((1, 3, 1))
-    query = c.text_input("", value="clouds at sunset")
+    query = c.text_input("", value="koala")
     corpus = st.radio("", ["Unsplash", "Movies"])
 
     if len(query) > 0:
         if "/" in query:
             queries = query.split("/")
-            clip_query, owl_query = ("/").join(queries[:-1]), queries[-1]
+            clip_query, owl_query = ("/").join(queries[:-1]).strip(), queries[
+                -1
+            ].strip()
         else:
             clip_query, owl_query = query, query
         retrieved = image_search(clip_query, corpus)