That was great; Flux is obviously a cutting-edge model that outperforms SD in so many ways; most specifically with composition, text, finger-counts, etc.
yea, once there's some good loras made for it and it doesn't look so surreal plastic it should be pretty rad. super excited about flux. it currently kinda has this midjourney stylized thing going on with it.
Before I throw money down, how is this for pixel art, specifically consistent art at different angles? Currently Midjourney is terrible and only Dalle-3 is able to output anything of quality.
Is there an option to upload, or use loras when generating an image? I'm super interested in being able to use img-img with flux on my base images, but I like to create specific videogame or anime characters, so loras are almost essential.
Two hands forming a heart shape against a plain background. The hands are coming together from the left and right sides of the image, with fingers slightly curved. The space between the hands creates a clear heart silhouette in the center. Soft, warm lighting emphasizes the contours of the hands. The image should have a tender, romantic mood. Photorealistic style." claude
"Two hands making the shape of a heart with the fingers" - got it first time using Fooocus. That whole thing about the red lips is probably throwing the generator off.
Yeah, its just better to run your ai locally. Far cheaper. You can run a version of flux locally, but that comes with its own issues. The rough part is saving up for over a year to grab a decent GPU. I'm currently limited to 12 gigs of VRAM, and while i can create amazing images and video, the TIME consumption is a beast. there are upsides to running locally and upsides to using someones paid service for ai.
It's always nice to see my generated images featured in the videos. ❤
Keep creating awesome images :)
there's no "my" in ai generated images lol
That was great; Flux is obviously a cutting-edge model that outperforms SD in so many ways; most specifically with composition, text, finger-counts, etc.
yea, once there's some good loras made for it and it doesn't look so surreal plastic it should be pretty rad. super excited about flux. it currently kinda has this midjourney stylized thing going on with it.
A comparison between Pro, Dev and Schnell would be interesting.😊
You got it!
It seems like almost all of the prompts used here were tuned for Flux?
Test it out yourself. It isn’t even close. Flux crushes SD3
Flux Pro vs SD3 Medium? Why not SD3 Ultra? What a fair comparison?
Sd3 ultra is twice the cost per image.
Oh youre not running either of these locally, are ya. =/
Was nice to see a remix of one of my silver dust Lora car featured in the videos. I'm addicted to the platform in a good way lol
Keep creating amazing art! Thank you so much for being part of the site, and for showing off your creations
Rather than comparing it to SD3, it might be better to compare it to closer competitors like Ideogram or the Phoenix model in Leonardo AI
Could you please release a tutorial on how to custom train flux for custom output? Your video on SDXL lora was really good
Before I throw money down, how is this for pixel art, specifically consistent art at different angles? Currently Midjourney is terrible and only Dalle-3 is able to output anything of quality.
Head to my discord and send me some examples
It doesnt know pixelart. Maybe with a lora in the future
We want a detailed comparison of flux pro with mj v6.1..
Flux absolutely destroyed SD3. Not even close. So excited for the open-source community to train some checkpoints and LoRAs on Flux.
I can’t wait to see what people create!
Can these models work similar to diffuser pipelines, like without UI?
great vid! are these better than Fooocus? thank you
It’s different. This has 20+ tools all in one system where as fooocus is just about image generation. Pixel dojo just does significantly more
Is there an option to upload, or use loras when generating an image?
I'm super interested in being able to use img-img with flux on my base images, but I like to create specific videogame or anime characters, so loras are almost essential.
Hi Brian, nice one ... do you plan adding a face swap tool to pixeldojo anytime soon?
Hi, I'm going to buy the subscription Can I ask,
I get to use Photos I create for commercial purposes?
And the Upscaler also unlimited?
Yes, and yes.
Two hands forming a heart shape against a plain background. The hands are coming together from the left and right sides of the image, with fingers slightly curved. The space between the hands creates a clear heart silhouette in the center. Soft, warm lighting emphasizes the contours of the hands. The image should have a tender, romantic mood. Photorealistic style."
claude
Hi there...question. Are the Lora's on pixedojo trained with Flux? thanks
Soon. Cost is currently too high
@@allyourtechai ok...thanks
"Two hands making the shape of a heart with the fingers" - got it first time using Fooocus. That whole thing about the red lips is probably throwing the generator off.
I’ll show you the original image. It’s complicated
I'll consider it if it was 10$ per month
That wouldn’t cover my expenses, sorry
@@allyourtechai but I would be an extra 10$ better than not having me at all, life is hard and that's what I can afford monthly.
Yeah, its just better to run your ai locally. Far cheaper. You can run a version of flux locally, but that comes with its own issues.
The rough part is saving up for over a year to grab a decent GPU.
I'm currently limited to 12 gigs of VRAM, and while i can create amazing images and video, the TIME consumption is a beast.
there are upsides to running locally and upsides to using someones paid service for ai.
And absolutely no frame of the video is actually "side by side".