I2V and the latest https://github.com/feizc/CogvideX-Interpolation models which is a fine-tune of the official I2V are just better at coherent motion provided a frame. However, these models have significant limitations with resolution and frame length... things CogVideoX-Fun solves.
What is the process of creating a "fun" version of a CogVideoX model like Interpolation? May you guys please consider doing a "Fun" version of the https://github.com/feizc/CogvideX-Interpolation finetune please?
I2V and the latest https://github.com/feizc/CogvideX-Interpolation models which is a fine-tune of the official I2V are just better at coherent motion provided a frame. However, these models have significant limitations with resolution and frame length... things CogVideoX-Fun solves.
What is the process of creating a "fun" version of a CogVideoX model like Interpolation? May you guys please consider doing a "Fun" version of the https://github.com/feizc/CogvideX-Interpolation finetune please?