Check out our Free Browser Extension that helps you with writing prompts for AI image generators 👇 chromewebstore.google.com/detail/prompt-catalyst/hehieakgdbakdajfpekgmfckplcjmgcf
No bias, who is better at generating what u want the best way posible with the greatest detail: Mid journey. Prove me wrong or tell me which other can be better
@@yubing0625I mean one is better because it's hard trained SD is not trained model the community does the training that the main point of stable diffusion
@@yubing0625 i thought the whole point was it looks better out the box, but with the "right prompts /negative prompts & data model installed" you can get as good if not better results
Check out our Free Browser Extension that helps you with writing prompts for AI image generators 👇
chromewebstore.google.com/detail/prompt-catalyst/hehieakgdbakdajfpekgmfckplcjmgcf
you should compare the content policy how much does it wimp out like chatgpt?
what model are you using? sd3 8b (the one on site, its not trained properly yet) or sd3 2b (the one open sourced , totally messed up)
Open source Stable Diffusion 3 Medium from Hugging Face
@@ai.catalyst oh thats why its so burned, the model is very bad that sai is retraining it and will release a 3.1 soon
midjourney has a distinct colour grading.
pretty good comparision
Midjourney V6 and Stable Diffusion 3 are both solid choices. I’ve been using CGDream lately and it’s been a amazing for my projects
Like it or not, Midjourney v6 is just better
No bias, who is better at generating what u want the best way posible with the greatest detail: Mid journey. Prove me wrong or tell me which other can be better
This video was sponsored by Midjourney*
I wish 😂
Midjourney is for noobs, Stable Diffusion is for pros
Midjourney looks better ngl. At least here
Midjourney obviously looks better, but Stable Diffusion is free and isn’t as restricted. I hope SD will catch up.
@@yubing0625I mean one is better because it's hard trained SD is not trained model the community does the training that the main point of stable diffusion
@@yubing0625 i thought the whole point was it looks better out the box, but with the "right prompts /negative prompts & data model installed" you can get as good if not better results