Update app.py

app.py

@@ -11,27 +11,8 @@ import gradio as gr
 
 # Load the open CLIP model
 device = "cuda" if torch.cuda.is_available() else "cpu"
-model, preprocess = clip.load("ViT-B/32", device=device)
-
-
-
-def search_video(search_query, display_heatmap=True, display_results_count=1):
-  
-    # Encode and normalize the search query using CLIP
-    with torch.no_grad():
-      text_features = model.encode_text(clip.tokenize(search_query).to(device))
-      text_features /= text_features.norm(dim=-1, keepdim=True)
-  
-    # Compute the similarity between the search query and each frame using the Cosine similarity
-    similarities = (100.0 * video_features @ text_features.T)
-    values, best_photo_idx = similarities.topk(display_results_count, dim=0)
-  
-  
-    for frame_id in best_photo_idx:
-      frame = video_frames[frame_id]
-      # Find the timestamp in the video and display it
-      seconds = round(frame_id.cpu().numpy()[0] * N / fps)
-    return frame,f"Found at {str(datetime.timedelta(seconds=seconds))}"
+model, preprocess = clip.load("ViT-B/32", device=device)  
+   
      
 
 def inference(video, text):
@@ -81,7 +62,25 @@ def inference(video, text):
   # Print some stats
   print(f"Features: {video_features.shape}")
  
-  return search_video(text)
+ 
+ search_query=text
+ display_heatmap=False
+ display_results_count=1
+  # Encode and normalize the search query using CLIP
+  with torch.no_grad():
+    text_features = model.encode_text(clip.tokenize(search_query).to(device))
+    text_features /= text_features.norm(dim=-1, keepdim=True)
+
+  # Compute the similarity between the search query and each frame using the Cosine similarity
+  similarities = (100.0 * video_features @ text_features.T)
+  values, best_photo_idx = similarities.topk(display_results_count, dim=0)
+
+
+  for frame_id in best_photo_idx:
+    frame = video_frames[frame_id]
+    # Find the timestamp in the video and display it
+    seconds = round(frame_id.cpu().numpy()[0] * N / fps)
+  return frame,f"Found at {str(datetime.timedelta(seconds=seconds))}"
   
 title = "Video Search"
 description = "demo for Anime2Sketch. To use it, simply upload your image, or click one of the examples to load them. Read more at the links below."