v3 Fast Video Lip Sync Change Lip Sync Video Change Audio Change Voiceover Change Lip Sync Change Voice

This video requires the face to remain in the video at all times (and the mouth must not be obstructed by any objects), otherwise an error will occur: face not detected (in this case, you need to crop out the parts without a face). If you don't want to crop, you can use this program for lip-syncing (https://www.runninghub.cn/ai detail/1977927668878430210).

Fan benefits: Click the avatar in the upper right corner, enter the invitation code in the Invitation Code section, and receive 1000 RH coins. Logging in daily earns an additional 100 coins, allowing you to generate many images and videos for free!

Invitation code: s0cwskom
For questions, you can add WeChat: ai_88183

This algorithm performs well, is less prone to GPU errors (out of memory), less likely to distort, and runs relatively fast.

A 48G graphics card (PLUS) can generate 2 seconds of video in roughly one minute. If your (membership single task duration) is 40 minutes or 60 minutes, you can generate 30 or 100 seconds of video at once. 

Users can set videos to be watermark-free by going to the Member Center — scroll down to find Watermark Management — Watermark Settings — No Watermark.

Popular image and video materials shared: https://pan.quark.cn/s/473668d9d78c
ComfyUI model download, Quark Cloud: https://pan.quark.cn/s/03c9a2c73166 (Mobile users can save to receive 1TB of free cloud storage from the official platform).
ComfyUI model download, Baidu Cloud shared file: https://pan.baidu.com/s/1QfG iKZ0c7eysCfgkIwuaw?pwd=7nup Extraction code: 7nup
(ComfyUI integration package) Quark Cloud shared file: https://pan.quark.cn/s/b84394c36c8d

Usage instructions (parameters pointed by arrows are modifiable):