找回密码
 加入我们
搜索
      
楼主: gartour

[软件] SVP手动更新RIFE最新模型教程

[复制链接]
发表于 2024-3-28 04:33 | 显示全部楼层
gartour 发表于 2024-3-25 23:54
是说lite模型吧。

今天出4.16 lite了

说的是v2版本,每个模型都有v2版本的。4.16lite也有v2版的。
发表于 2024-4-4 13:05 | 显示全部楼层
哪位能说明一下,lite和v2?  lite肯定是性能要好一些,这个试试就能看出来,这个v2版是有啥讲究?
 楼主| 发表于 2024-4-4 14:41 | 显示全部楼层
netlzh 发表于 2024-4-4 13:05
哪位能说明一下,lite和v2?  lite肯定是性能要好一些,这个试试就能看出来,这个v2版是有啥讲究? ...

"RIFE v2 models which handle paddings internally and reduce memory transactions on heterogeneous devices."

简单说v2节省显存带宽和pcie带宽。
发表于 2024-4-5 11:58 | 显示全部楼层
gartour 发表于 2024-4-4 14:41
"RIFE v2 models which handle paddings internally and reduce memory transactions on heterogeneous d ...

看来用 v2 的 lite 版本肯定是没有错的了。
发表于 2024-4-20 11:49 | 显示全部楼层
为啥我都更新到最新版,同样选不了RIFE模型呢
屏幕截图 2024-04-20 114659.png
屏幕截图 2024-04-20 114931.png
 楼主| 发表于 2024-4-20 14:16 | 显示全部楼层
小枫肥刀 发表于 2024-4-20 11:49
为啥我都更新到最新版,同样选不了RIFE模型呢

路径错误
发表于 2024-4-20 15:15 | 显示全部楼层
gartour 发表于 2024-4-4 14:41
"RIFE v2 models which handle paddings internally and reduce memory transactions on heterogeneous d ...

双卡X8这种可能就好一点了

比如我一张卡跑RIFE  一张卡跑AnimeJanaiV3
发表于 2024-4-30 19:04 来自手机 | 显示全部楼层
为啥4.16只有lite?
发表于 2024-5-13 00:50 | 显示全部楼层
strike94 发表于 2023-12-10 23:24
4K 倍帧 ,使用4.13V2的模型,4090都跑不动。。真的裂开

你这种应该是着色器开太猛了
试着清空着色器,应该就好多了
然后找个合适的着色器组合
发表于 2025-7-19 06:19 | 显示全部楼层
不好意思,挖个lz的老帖
在主页说明有这么一个段落:
vsort: ONNX Runtime-based CPU/GPU Runtime
ONNX Runtime is an AI inference runtime with many backends.

The vs-onnxruntime plugin provides optimized CPU and CUDA GPU runtime for some popular AI filters.

To install, download the latest release and extract them into your VS plugins directory.

Please visit the vsort directory for details.

vstrt: TensorRT-based GPU Runtime
TensorRT is a highly optimized AI inference runtime for NVidia GPUs. It uses benchmarking to find the optimal kernel to use for your specific GPU, and so there is an extra step to build an engine from ONNX network on the machine you are going to use the vstrt filter, and this extra step makes deploying models a little harder than the other runtimes. However, the resulting performance is also typically much much better than the CUDA backend of vsort.

TensorRT-RTX is a specialization of TensorRT for NVIDIA RTX GPUs, which compiles engines faster with comparable performance with TensorRT.

To install, download the latest release and extract them into your VS plugins directory.

Please visit the vstrt directory for details.

vsmigx: MIGraphX-based GPU Runtime
MIGraphX is a highly optimized AI inference runtime for AMD GPUs. It also uses benchmarking to find the optimal kernel, similar to vstrt.

To install, download the latest release and extract them into your VS plugins directory.

Please visit the vsmigx directory for details.


意思是VapourSynth内的文件也要一起更新吗?
 楼主| 发表于 2025-7-19 16:45 | 显示全部楼层
本帖最后由 gartour 于 2025-7-19 16:47 编辑
ljy1414 发表于 2025-7-19 06:19
不好意思,挖个lz的老帖
在主页说明有这么一个段落:


https://github.com/AmusementClub/vs-mlrt

NVidia GPU: vsort-cuda, vstrt & vstrt_rtx, vsncnn-vk

svp用了这几个包里面的部分文件,我曾经是找出来覆盖更新了的(务必先备份)。具体哪些文件我没记,对比一下花不了多少功夫。

不过svp不久前有一次大更新,这些runtime是一并更新了的,如果你是正版同步更新的,不太必要手动去弄了。
发表于 2025-7-20 01:59 | 显示全部楼层
gartour 发表于 2025-7-19 16:45
https://github.com/AmusementClub/vs-mlrt

NVidia GPU: vsort-cuda, vstrt & vstrt_rtx, vsncnn-vk

感谢帮忙,实测了一下
只需要更新scripts、VSTRT-Windows-x64这两个压缩包
您需要登录后才可以回帖 登录 | 加入我们

本版积分规则

Archiver|手机版|小黑屋|Chiphell ( 沪ICP备12027953号-5 )沪公网备310112100042806 上海市互联网违法与不良信息举报中心

GMT+8, 2025-7-22 15:43 , Processed in 0.010571 second(s), 5 queries , Gzip On, Redis On.

Powered by Discuz! X3.5 Licensed

© 2007-2024 Chiphell.com All rights reserved.

快速回复 返回顶部 返回列表