Search Results
7/19/2025, 1:59:16 PM
>>937307966
In the typical wan workflows all it does is generate a series of images, you can do whatever you want to them before saving them as a video.
My workflow basically takes an image, downscales it to a set number of megapixels and feeds that to Wan. It also takes my prompt and the image and feeds it into an LLM for a more flowery description. Those are fed into the positive. The series of images from Wan are then upscaled back to the original res, interpolated to 3x the framerate, and saved as a video. There are also intermediate videos saved and a downscaled version for posting here.
>>937307967
<3
In the typical wan workflows all it does is generate a series of images, you can do whatever you want to them before saving them as a video.
My workflow basically takes an image, downscales it to a set number of megapixels and feeds that to Wan. It also takes my prompt and the image and feeds it into an LLM for a more flowery description. Those are fed into the positive. The series of images from Wan are then upscaled back to the original res, interpolated to 3x the framerate, and saved as a video. There are also intermediate videos saved and a downscaled version for posting here.
>>937307967
<3
Page 1