|
|
@@ -27,10 +27,12 @@ Bo Dai
|
|
|
## Setup for Inference
|
|
|
|
|
|
### Prepare Environment
|
|
|
-Our approach takes around 60 GB GPU memory to inference. NVIDIA A100 is recommanded.
|
|
|
+~~Our approach takes around 60 GB GPU memory to inference. NVIDIA A100 is recommanded.~~
|
|
|
+
|
|
|
+***We updated our inference code with xformers and a sequential decoding trick. Now AnimateDiff takes only ~12GB VRAM to inference, and can be run on a single RTX3090 !!***
|
|
|
|
|
|
```
|
|
|
-git clone https://github.com/guoyww/animatediff.git
|
|
|
+git clone https://github.com/guoyww/AnimateDiff.git
|
|
|
cd AnimateDiff
|
|
|
|
|
|
conda env create -f environment.yaml
|
|
|
@@ -107,7 +109,7 @@ Here we demonstrate several best results we found in our experiments or generate
|
|
|
<td><img src="__assets__/animations/model_07/04.gif"></td>
|
|
|
</tr>
|
|
|
</table>
|
|
|
-<p style="margin-left: 2em; margin-top: -1em">Model:<a href="https://civitai.com/models/107295?modelVersionId=115371">holding_sign</a> (samples are contributed by CivitAI artists)</p>
|
|
|
+<p style="margin-left: 2em; margin-top: -1em">Model:<a href="https://civitai.com/models/107295?modelVersionId=115371">holding_sign</a>, etc. (samples are contributed by CivitAI artists)</p>
|
|
|
|
|
|
<table class="center">
|
|
|
<tr>
|