Update app.py

app.py

@@ -55,7 +55,7 @@ d = 1024
 text_index = faiss.index_factory(d, "Flat", faiss.METRIC_INNER_PRODUCT)
 faiss.read_index("text_index.index")
 
-def T2Isearch(query,focussed_word, k=50):
+def T2Isearch(query,focussed_word, k=5):
     # Encode the text query        
     inputs = text_tokenizer([query,focussed_word], padding=True, return_tensors="pt")    
     outputs = text_model(**inputs)
@@ -63,22 +63,25 @@ def T2Isearch(query,focussed_word, k=50):
     query_vector = query_embedding.detach().numpy()
     query_vector = np.concatenate((query_vector[0], query_vector[1]), dtype=np.float32)
     query_vector = query_vector.reshape(1,1024)
-    # query_vector = test_xt[0]
-    # query_vector = np.array([query_embedding])
     faiss.normalize_L2(query_vector)
-    # text_index.nprobe = index.ntotal
-    text_index.nprobe = 100
+    text_index.nprobe = text_index.ntotal
+
+    # text_index.nprobe = 100
     
     # Search for the nearest neighbors in the FAISS text index
     D, I = text_index.search(query_vector, k)
+    print(D)
+    print(I)
 
     # get rank of all classes wrt to query
     classes_all = []
     Y = train_yt
     neighbor_ys = Y[I]
     class_freq = np.zeros(Y.shape[1])
+    
     for neighbor_y in neighbor_ys:
         classes = np.where(neighbor_y > 0.5)[0]
+        print(classes)
         for _class in classes:
             class_freq[_class] += 1
 
@@ -91,6 +94,7 @@ def T2Isearch(query,focussed_word, k=50):
 
     lis  = ['aeroplane', 'bicycle','bird','boat','bottle','bus','car','cat','chair','cow','diningtable','dog','horse','motorbike','person','pottedplant','sheep','sofa','train','tvmonitor']
     class_ = lis[ranked_classes_after_knn[0]-1]
+    print(class_)
 
     # Map the image ids to the corresponding image URLs
     for i in range(len(image_list)):