←
Jukka Seppänen1,780 posts
Jukka Seppänen avatar

Jukka Seppänen

@Kijaidesign

3D modeling/printing artist, AI enthusiast, rookie Python coder.

◷ Joined October 2017▧ 753 Photos and videos
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·4 Mar 2025

WanVideo2video fun, 129 frames with sliding context windows using the 14B model.

0:08
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·3 Mar 2025

WanVideo sliding context window test on the 480p 14B T2V -model, worked surprisingly well but incredibly slow (50 mins on a 5090, 513 frames at 832x480)

0:32
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·2 Mar 2025

As an impatient person, I often get obsessed with optimizations, tried implementing TeaCache for WanVideo, failed, but in the process accidentally (maybe) succeeded? Results look promising at least!

0:03
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·1 Mar 2025

WanVideo has been lot of fun, currently playing with applying sliding context windowing similar to AnimateDiff, the 1.3B model is especially suitable for this due to it's speed, here's 1025 frames in one go, under 5GB VRAM used, but took about 10 mins on a 5090 with 30 steps.

1:04
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·4 Dec 2024

Couple more examples, just changing prompt.

0:04
0:04
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·4 Dec 2024

HunhuyanVideo -model's vid2vid passes the hippo test! Very promising and versatile model, thanks to the cfg distillation and every possible optimization I could think of this clip of 101 frames at 768x432 took about 2 minutes to sample, fitting slightly under 20GB using my nodes.

0:04
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·19 Nov 2024

I have finally pushed a bigger update to my CogVideoX ComfyUI wrapper nodes, cleaning up most of the bloat that has been accumulating with all these different models. One of the discoveries I made during this is that the orbit -LoRAs work with the "Fun" -models as well!

0:08
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·6 Oct 2024

I've been deep into CogVideoX -code lately. Using the CogVideoX-Fun "pose" model allows control with variety of inputs, and applying Kosinkadink's context windowing from AnimateDiff (thanks to ) I've managed some decently stable longer generations. #ai #cogvideox

0:09
0:06
Jukka Seppänen avatar
Jukka Seppänen@Kijaidesign·7 Jan 2024

Example of what AI generation can do in practice currently, from initial image (which is also generated) to a very detailed printable model, the level of detail mostly made possible with the Marigold depth estimation. #ai #3dprinting #dragon #technology

0:37