This LoRA can generate a video that plays the image of your choice in reverse from the start frame by inputting a pure white image into the end frame of wanvideo_480p_I2V_endframe_example_01.json included in the example_workflows of kijai's WanVideoWrapper. For the I2V 720p model, using wanvideo_480p_I2V_endframe_example_01.json will give you the expected results. The model was created by training a time-lapse video of a person drawing , using only two videos. Please do not use this LoRA for malicious purposes.
Description
FAQ
Comments (8)
"Please do not use this LoRA for malicious purposes." I wonder what you do you mean. Is it NSFW material?
I think more like trying to confuse people by claiming an image you generated is not actually AI and giving that video as a proof.
@TroublesomeA Oh that makes sense
Could you please share the data preparation process and training scripts you used?
https://x.com/8co28/status/1904156266194260037
The dataset is these two videos. The training script is the default one from musubi-tuner.
wow this is really cool
我也打开了这个工作流 但是我不知道在什么地方加载这个LORA....
need this on google colab!!!
Details
Files
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.