Same as above, but increasing the resolution to 1280x720. These are using the FusionX merge.

Testing the 14B Wallace & Gromit Wan LoRA made by @ChetiArt with the obligatory Puerto Rican bar dinosaur guitar prompt 🦖 #aivideo

Testing the wave motion from above with other images 🌊

You can try out Particle Grid Animator here ✨.

Sharing a Particle Grid Animation tool to create tracks like this to drive image animation in ComfyUI using Wan and ATI. Really interesting results! 👇

I was curious what would happen if the hands crossed over each other, so tried another version with a few seeds. New input at the top left and results of three different seeds for comparison. Seed 1 is what I had in mind, not 2 or 3! 🫸🫷

AE + Comfy process (and links to resources) here:

Sharing about using After Effects keyframes to animate a still image with Wan and ATI in ComfyUI 👇

Chattable Wan Video knowledge archive built by scraping chats in the @banodoco discord. Try it out! 🔗

I can now publicly share my chattable Wan AI Video Knowledge Base on NotebookLM. You can use either voice or text chat. A little silly, but here's an example of me talking to the podcast hosts to catch up on the last week of news. Link below!

There also seems to be quite a lot of graininess and noise in some images. (see result on the right) Anyone know what's up with that? I see the same thing Krea and Fal and with both the Pro and Max versions of Kontext.

Trying out photo editing with FLUX Kontext. It really keeps the subject consistent!. Prompt structure here is "Turn the boy into a [marble] sculpture, keep the background the same"

Here's how I'm splitting the clip and making the mask (total of 81 frames) to feed into a Wan Video VACE Encode node. Thanks to @noma_door for the inspiration! 🙏

Looping a video with Wan + VACE in ComfyUI. This uses a 30 frame chunk of the source video, splits it into two 15 frame parts to loop, then inpaints the middle! It blends into the source motion this way. ♾. (picture of the nodes/setup in thread) #aivideo

The basic idea is to set the style using ChatGPT, but then @Magnific_AI Mystic actually matches it for use as a first frame in Comfy and brings it back to 16:9 aspect.

Meanwhile in the open-source end of the pool, here's Wan + VACE + a style frame in ComfyUI 🤘 #aivideo . Process images in thread 🧵

This second one is giving Mountain Goats for sure.

Here's another one it made!

Wow. My dino prompt, but now in Veo 3 😳. "A dinosaur with a white fedora and a Hawaiian shirt playing an acoustic guitar on stage at a small waterside bar in Puerto Rico. The dinosaur is singing about how thing used to be in the Mesozoic Era. People are clapping and laughing"

Also! Just took this a step further and converted the channel into something I can chat with using NotebookLM. Process: