Merge pull request #500 from THUDM/main

Merge

Author: zRzRzRzRzRzRzR

README.md

@@ -189,7 +189,7 @@ models we currently offer, along with their foundational information.
   <tr>
     <td style="text-align: center;">Video Resolution</td>
     <td colspan="1" style="text-align: center;">1360 * 768</td>
-    <td colspan="1" style="text-align: center;">256 <= W <=1360<br> 256 <= H <=768<br>  W,H % 16 == 0</td>
+    <td colspan="1" style="text-align: center;"> Min(W, H) = 768 <br> 768 ≤ Max(W, H) ≤ 1360 <br> Max(W, H) % 16 = 0 </td>
     <td colspan="3" style="text-align: center;">720 * 480</td>
   </tr>
   <tr>

README_ja.md

@@ -181,7 +181,7 @@ CogVideoXは、[清影](https://chatglm.cn/video?fr=osm_cogvideox) と同源の
   <tr>
     <td style="text-align: center;">ビデオ解像度</td>
     <td colspan="1" style="text-align: center;">1360 * 768</td>
-    <td colspan="1" style="text-align: center;">256 <= W <=1360<br> 256 <= H <=768<br>  W,H % 16 == 0</td>
+    <td colspan="1" style="text-align: center;"> Min(W, H) = 768 <br> 768 ≤ Max(W, H) ≤ 1360 <br> Max(W, H) % 16 = 0 </td>
     <td colspan="3" style="text-align: center;">720 * 480</td>
   </tr>
   <tr>

README_zh.md

@@ -171,7 +171,7 @@ CogVideoX是 [清影](https://chatglm.cn/video?fr=osm_cogvideox) 同源的开源
   <tr>
     <td style="text-align: center;">视频分辨率</td>
     <td colspan="1" style="text-align: center;">1360 * 768</td>
-    <td colspan="1" style="text-align: center;">256 <= W <=1360<br> 256 <= H <=768<br>  W,H % 16 == 0</td>
+    <td colspan="1" style="text-align: center;"> Min(W, H) = 768 <br> 768 ≤ Max(W, H) ≤ 1360 <br> Max(W, H) % 16 = 0 </td>
     <td colspan="3" style="text-align: center;">720 * 480</td>
     </tr>
   <tr>

tools/parallel_inference/parallel_inference_xdit.py

@@ -78,8 +78,9 @@ def main():
         num_frames=input_config.num_frames,
         prompt=input_config.prompt,
         num_inference_steps=input_config.num_inference_steps,
-        generator=torch.Generator(device="cuda").manual_seed(input_config.seed),
+        generator=torch.Generator().manual_seed(input_config.seed),
         guidance_scale=6,
+        use_dynamic_cfg=True,
     ).frames[0]
 
     end_time = time.time()