the program that created consistent characters in the demo is not dalle 3, it was a feature of gpt-4o that is not released yet just like the voice feature.
@aiconcoction
3 ай бұрын
Thank you.
@halilcengiz1983
2 ай бұрын
teşekkürler, faydası oldu benim için. ben bir haftadır aynı karakterleri oluşturmaya çalışıyorum ama zorlanıyorum. bu teknikleride deneyeceğim. 20 dolar vermemize ragmen çok kolay olmuyor
@aiconcoction
2 ай бұрын
Rica ederim Bazen istenen sonuçları elde etmek için birçok nesil ve revizyon gerekir. Umarım OpenAI tutarlı karakter özelliğini yakında yayınlar.
@saoitb1729
3 ай бұрын
the update came may 13th, but things like voice and character consistency still havent rolled out, that why there is no cosistant videos although gtp says so
@aiconcoction
3 ай бұрын
Thank you.
@sylv1montreuil
3 ай бұрын
Great video. I am having similar issues. Doesn’t work in Dalle3 api nor in GPT4o. Does anyone knows when this feature will be rolled out on Dalle3 API?
@aiconcoction
2 ай бұрын
Thank you. I don't know when it will be rolled out, but I hope it's soon.
@ToysFroMali
Ай бұрын
Hello. I'm a designer and I'm looking for information about AI that can help me with my work. I'm looking for an AI that can be trained to make a specific character. Like a drawing of mine in just one style and pose and that it can recreate the character keeping the same layout, lines, features and clothing, but in different positions. However, so far I haven't found any. They all change the character completely, including the style of the drawing. However, it's becoming a challenge. Do you know of any AI that is like this and that is effective in what I'm looking for?
@aiconcoction
Ай бұрын
AI models often struggle to maintain exact consistency across images due to their tendency to interpret prompts creatively. However, here are some strategies you can try: 1. Use tools that allow you to train your own models, such as Leonardo, TensorArt, Runway ML, and Stable Diffusion. 2. Utilize tools that support style references, image-to-image, and pose-to-image techniques. Leonardo is effective in this regard, and MidJourney also allows for style references. 3. Edit your images using an image editor to correct minor inconsistencies.
@Agente00Ai
3 ай бұрын
Yes. Free accounts have access to 4o, as it rolls out. No generative art available
@aiconcoction
3 ай бұрын
Thank you.
@pomega
2 ай бұрын
Its tough cause its not listening me! For 2d design it sucks
@aiconcoction
2 ай бұрын
I know, it took me many generations to get the consistent images for the video, except for the mail woman.
Пікірлер: 17