KMeans index extraction actually works and creates smaller index files.

Author: AznamirWoW

rvc/train/process/extract_index.py

@@ -46,9 +46,7 @@
     np.random.shuffle(big_npy_idx)
     big_npy = big_npy[big_npy_idx]
 
-    if big_npy.shape[0] > 2e5 and (
-        index_algorithm == "Auto" or index_algorithm == "KMeans"
-    ):
+    if big_npy.shape[0] > 2e5 or index_algorithm == "KMeans":
         big_npy = (
             MiniBatchKMeans(
                 n_clusters=10000,