|
|
发表于 2023-7-6 13:08:11
|
显示全部楼层
</pre>
This is the DirectML version of TG-Plus, used for testing in non-cuda environment.
<pre>
NOTE:1.Only works on JVP_112b+2.0.
2.It is very slow compared to cuda version.
3.Removed TecoGAN and DA model, other functions same as the standard version basically.
4.The DirectML version is a standalone and cannot be mixed with the cuda version.
5.If updates, it will be released separately.
PS:适用于没有CUDA的硬件设备;运行非常慢;不可和CUDA版本混合;去除了TecoGAN and DA模型 |
|