1115

README.md

@@ -41,8 +41,8 @@ Below is a table listing information on the video generation models available in
   <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>
-  </tr>
+    <td colspan="1" style="text-align: center;"> Min(W, H) = 768 <br> 768 ≤ Max(W, H) ≤ 1360 <br> Max(W, H) % 16 = 0 </td>
+    </tr>
   <tr>
     <td style="text-align: center;">Inference Precision</td>
     <td colspan="2" style="text-align: center;"><b>BF16 (recommended)</b>, FP16, FP32, FP8*, INT8, not supported INT4</td>
@@ -53,7 +53,7 @@ Below is a table listing information on the video generation models available in
   </tr>
   <tr>
     <td style="text-align: center;">Multi-GPU Inference Memory Consumption</td>
-    <td colspan="2" style="text-align: center;"><b>BF16: 15GB* </b><br></td>
+    <td colspan="2" style="text-align: center;"><b>BF16: 24GB* </b><br></td>
   </tr>
   <tr>
     <td style="text-align: center;">Inference Speed<br>(Step = 50, BF16)</td>
@@ -118,9 +118,9 @@ conversion to get a better experience.**
 1. Install the required dependencies
 
 ```shell
-# diffusers>=0.32.0
-# transformers>=0.46.2
-# accelerate>=1.0.1
+# diffusers>=0.32.0dev (or from source)
+# transformers>=4.46.2
+# accelerate>=1.1.1
 # imageio-ffmpeg>=0.5.1
 pip install --upgrade transformers accelerate diffusers imageio-ffmpeg
 ```

README_zh.md

@@ -27,7 +27,7 @@ CogVideoX是 [清影](https://chatglm.cn/video?fr=osm_cogvideo) 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>
     </tr>
   <tr>
     <td style="text-align: center;">推理精度</td>
@@ -39,7 +39,7 @@ CogVideoX是 [清影](https://chatglm.cn/video?fr=osm_cogvideo) 同源的开源
   </tr>
   <tr>
     <td style="text-align: center;">多GPU显存消耗</td>
-    <td colspan="2" style="text-align: center;"><b>BF16: 15GB* </b><br></td>
+    <td colspan="2" style="text-align: center;"><b>BF16: 24GB* </b><br></td>
   </tr>
   <tr>
     <td style="text-align: center;">推理速度<br>(Step = 50, BF16)</td>
@@ -98,9 +98,9 @@ pipe.vae.enable_tiling()
 1. 安装对应的依赖
 
 ```shell
-# diffusers>=0.32.0
-# transformers>=0.46.2
-# accelerate>=1.0.1
+# diffusers>=0.32.0dev (or from source)
+# transformers>=4.46.2
+# accelerate>=1.1.1
 # imageio-ffmpeg>=0.5.1
 pip install --upgrade transformers accelerate diffusers imageio-ffmpeg 
 ```

vae/config.json

@@ -35,5 +35,6 @@
     "CogVideoXUpBlock3D"
   ],
   "use_post_quant_conv": false,
-  "use_quant_conv": false
+  "use_quant_conv": false,
+  "invert_scale_latents": true
 }