Update README.md

README.md

@@ -67,6 +67,8 @@ same-resolution warmup**.
 
 ## Video Comparison
 
+<video controls src="https://huggingface.co/lightx2v/RelayVSR_X2/resolve/main/assets/benchmark/comparison_grid.mp4"></video>
+
 Many super-resolution methods can amplify the temporal flicker already present in
 AI-generated input videos. **RelayVSR can substantially alleviate this flicker**,
 producing more temporally consistent results.
@@ -77,8 +79,6 @@ higher computational cost and slower inference. To reduce this cost, we trained
 combines effective temporal-flicker suppression with **fast inference**, making it
 suited to upscaling AI-generated videos.
 
-<video controls src="https://huggingface.co/lightx2v/RelayVSR_X2/resolve/main/assets/benchmark/comparison_grid.mp4"></video>
-
 ## Model Files
 
 Download both checkpoints into the same model directory: