Update README.md

README.md

@@ -25,8 +25,8 @@ Comparison of VSR datasets. Our proposed ViSpeR dataset is larger in size compar
 | **VoxCeleb2**   | 124             | 42              | --              | --              |
 | **AVSpeech**    | 122             | 270             | --              | --              |
 | **ViSpeR (TedX)** | 192 (160k)    | 207 (151k)      | 49 (48k)        | 129 (143k)      |
-| **ViSpeR (Wild)** | 799 (548k)    | 851 (531k)      | 1152 (1.01M)    | 658 (593k)      |
-| **ViSpeR (full)** | 991 (709k)    | 1058 (683k)     | 1200 (1.06M)    | 787 (736k)      |
+| **ViSpeR (Wild)** | 680 (481k)    | 587 (383k)      | 1152 (1.01M)    | 658 (593k)      |
+| **ViSpeR (full)** | 872 (641k)    | 794 (534k)     | 1200 (1.06M)    | 787 (736k)      |
 
 
 ## Downloading the data: