2 Frame GIF

[vital spire]

Only GIF maker that somewhat works xD
Specialist in crafting dynamic 2-frame GIFs from images and descriptions.
https://chat.openai.com/g/g-CNrssrlML-2-frame-gif

[oblique arrow]

Sometimes it behaves strangely, but generally it does what it promises. Good job,
I'm making my own gpt, how did you get it to analyze the newly created image? Have you used any specific technique?
thank you

some examples 🙂

[vital spire]

Those are great examples!

Image generation part was bit of a nightmare, DALL-E never wants to fully listen. So for this GPT I made this GPT obey its limitations. For example you can never get the image to split consistently top and bottom or left and right. So I included hotkey features to split the image in either direction. Trying to create more than 2 frames was also really inconsistent, so went with a 2 frame solution

[vital spire]

I've changed the code for a more aligned GIFs. Give it a try!

[oblique arrow]

the tree was eassy but the ninja it took me 9 trys. [prompt: superkawaii chibi ninja dress up with red leader making a slash with his blade --ar: 2:1]
im pretty sure you know but just in case github -> ProfSynapse -> GIFAnimator, i put here may be it helps (uau i post a link to github and win a 5min ban ... )
again thanks for your work

[prompt: superkawaii chibi lizard kaiju throwing fire from his mouth --ar: 2:1]

[vital spire]

thank you for the github i’ll check it out!

[pseudo wing]

the code anlayzes for the middle pixel color

that's actually a great idea to anyalyze image

[oblique arrow]

i dont know if openai made changos to day, but i cant get gif. download error or click and nothing download = blank page

[vital spire]

im updating it rn

it should be good now. Its using a longer algorithm from the github repo you posted

[vital spire]

lmk if it works

[oblique arrow]

i reach my limit XD so i cant test till an hour.

[vital spire]

Tweaking this thing eats all mine too 🤣

[vital spire]

Did a round 2 and added to suggesteds xd

[oblique arrow]

Cool 👍

I'm getting file not found when try to downloads. ,may be IS some thing related to openai? Similar error on 4 frame gif prompt i show previously . 🥲

[vital spire]

same for me :/ hopefully its an openai issue i couldnt seem to fix it

[oblique arrow]

I have been trying other gif generators and the download works. When you ask chatgpt to list the sandbox files in your gpt, the gif does not appear compared to the others that it does. and... I no longer have quota left

[oblique arrow]

Today i can download but the gif cut its working oddly

[vital spire]

Yea I had changed it to detect frame borders, but messed it up, so I rolled it back to how it was initially

[vital spire]

New thread with 2.0 @oblique arrow it turned out so much better with double image generation and gen_id https://discord.com/channels/974519864045756446/1177881410720256010

[oblique arrow]

The new technique is very interesting. The generation of the first image makes it look more eye-catching and lively. The drawback arises with the second image, as the transition from one to the next is very difficult to maintain and jumps too much in elements such as the background or even the main character. Still, the results are quite intriguing. Here are some examples.

What kind of prompt do you suggest? I have been testing: style + main character + action.

[vital spire]

That's totally fine. I haven't really tested to see which kind of prompts give most consistent results.