Instructions:
This workflow is based on a remix model and an anime model for image-to-video generation, capable of maintaining high motion.
The workflow uses the Image Batch Extend With Overlap node to connect consecutive video segments for longer video output.
The last frame of each video segment is preserved, allowing manual continuation.
LoRAs can be added to this workflow and selected at your discretion.
This workflow was created early in my practice. It performs well with LoRAs, but consistency is relatively weak for long videos. I usually maintain consistency by applying face swap on the last frame. My other workflow, SVI, can generate 12-second long videos with significantly better results.https://civarchive.com/models/2352785?modelVersionId=2646245
You can test this workflow online for free first. If you find it useful, you may download it and run it locally:
Online workflow: https://www.runninghub.ai/post/2017554271281090561/?inviteCode=rh-v1443
New registrants receive 1000 points free, and daily logins give an additional 100 points.Running one image takes about 4 minutes and costs around 60 points.
Credits & Model Download Links:
DaSiWa WAN 2.2 I2V 14B：https://civarchive.com/models/1981116/dasiwa-wan-22-i2v-14b-or-lightspeed-or-safetensors?modelVersionId=2555640
Wan2.2-Remix：https://civarchive.com/models/2003153/wan22-remix-t2vandi2v?modelVersionId=2567309