Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

replicate.com

Post reply on HN
Print view View original

71–80 of 159 posts

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#71

Post by vunderba »

miohtama wrote: post #68
loufe wrote: post #52

Earlier quoted context omitted.

That is astoundingly good adherence to the description. I already liked and was impressed by Flux1 but that is perhaps the most impressive image generation I've ever seen.

Is it going be able to go head-to-head against Midjourney?

MJ is by far the worst model for complex prompt ADHERENCE, though it has excellent compositional quality.

Comparisons of similar prompt using Midjourney 6.1

https://imgur.com/a/WBnPl7I

Also, flux (schnell, dev) can be run on your local machine.

If you really want to use a paid service, Ideogram is probably the best one out there that balances quality with adherence. DALL-E 3 also has good adherence as well though the quality can sometimes be iffy, and it's very puritanical in terms of censorship.

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#73

Post by andersa »

vunderba wrote: post #50
pdntspa wrote: post #47

Earlier quoted context omitted.

I wonder if you can use Flux to generate the base image then img2img on SD1.4 to impart artistic style?

That's what a refiner is for in auto1111. Taking an image the last 10% and touching it up with an alternative model. I actually use flux to generate image for purposes of adherence , then pull it in as a canny/depth controlnet with more established models like realvis, unstableXL, etc.

That is an interesting idea, I somehow hadn't thought of using flux in a chain like that, thanks!

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#74

Post by doctorpangloss »

throwaway314155 wrote: post #57
liuliu wrote: post #12

Earlier quoted context omitted.

What do you mean? FLUX.1 prompts women or women faces just fine? Do you mean the skin texture is unrealistic or some other artifacts?

what they really mean is that it's not useful for generating lewd imagery of women. It was likely nerfed in this regard on purpose because BFL didn't want to be associated with that (however legal it may be).

I'm not sure why you're being downvoted because I think this is a misconception that's worth clearing up. There is no aspect of what I'm doing that is lewd or lewd adjacent. I just want control of a character's face for making art for an open source game. While I do not totally understand what specific decisions Flux made that would make their model weak in the regard of specifying the appearance of someone's face, one thing is clear: the humanities people are right, this is like a great example of how censorship and Big Prude has impacted artmaking.

It is actually making it harder to use the technology to represent women characters, which is so ironic. That said, I could just lEaRn tO dRaW or pAy aN aRtIsT right? The discourse around this is so shitty.

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#75

Post by skort »

vessenes wrote: post #28

Flux is so frustrating to me. Really good prompt adherence, strong ability to keep track of multiple parts of a scene, it's technically very impressive. However it seems to have had no training on art-art. I can't get it to generate even something that looks like Degas, for instance. And, I can't even fine tune a painterly art style of any sort into Flux dev. I get that there was working, living artist backlash at SD…

>but, it's a real loss. Both in terms of human knowledge of, say composition, emotion, and so on, but also for style diversity

But that real art still exists, and can still be found, so what exactly is the loss here?

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#76

Post by drcongo »

zamadatix wrote: post #66
drcongo wrote: post #60

Earlier quoted context omitted.

Really? I tried using it in ComfyUI on my Mac Studio, failed, went searching for answers and all I could find said that something something fp8 can't run on a Mac, so I moved on.

If you're looking for a prebuilt "no tinkering" solution https://diffusionbee.com/ is an open source app (Github link at the bottom of the page if you want to see the code) which has a built in button to import Flux models at the bottom of the home screen.

Thanks, I'll take a look.

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#77

Post by waffletower »

zamadatix wrote: post #61
stoobs wrote: post #42

Earlier quoted context omitted.

I'm running Flux dev fine on a 3080 10GB, unquantised, on windows the nvidia drivers have a function to let it spill over into system ram. It runs a little slower, but it's not a deal-breaker unlike nvidia's pricing and power requirements at the moment

What are you using to run it? When I run Flux Dev in Windows using comfy on a 4090 (24 GB) sometimes it all crashes because it runs out of VRAM when I'm doing too much other stuff.

Not a good reference for windows -- I use HuggingFace APIs on cog/docker deployments in Linux. I needed to use `PYTORCH_NO_CUDA_MEMORY_CACHING=1 -e PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True` envvars to eliminate memory errors on the 3090s. When I run on the Mac there is enough memory not to require shenanigans. Runs approximately as fast as the 3090s but the 3090s heat my basement and the Mac heats my face.

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#78

Post by doctorpangloss »

thomastjeffery wrote: post #51
vessenes wrote: post #28

Flux is so frustrating to me. Really good prompt adherence, strong ability to keep track of multiple parts of a scene, it's technically very impressive. However it seems to have had no training on art-art. I can't get it to generate even something that looks like Degas, for instance. And, I can't even fine tune a painterly art style of any sort into Flux dev. I get that there was working, living artist backlash at SD…

I think that's part of what makes FLUX.1 so good: the content it's trained on is very similar . Diversity is a double-edged sword. It's a desirable feature where you want it, and an undesirable feature everywhere else. If you want an impressionist painting, then it's good to have Monet and Degas in the training corpus. On the other hand, if you want a photograph of water lilies, then it's good to keep Monet out of th…

DALL-E3 doesn't struggle with this. It's just opinions. There's no technical limitation. They chose to weaken the model in this regard.

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#79

Post by vessenes »

vunderba wrote: post #50
pdntspa wrote: post #47

Earlier quoted context omitted.

I wonder if you can use Flux to generate the base image then img2img on SD1.4 to impart artistic style?

That's what a refiner is for in auto1111. Taking an image the last 10% and touching it up with an alternative model. I actually use flux to generate image for purposes of adherence , then pull it in as a canny/depth controlnet with more established models like realvis, unstableXL, etc.

Yes, that is my current workflow as well.

Re: FLUX1.1 [pro] – New SotA text-to-image model from Black Forest Labs

#80

Post by vessenes »

crystal_revenge wrote: post #49
vessenes wrote: post #28

Flux is so frustrating to me. Really good prompt adherence, strong ability to keep track of multiple parts of a scene, it's technically very impressive. However it seems to have had no training on art-art. I can't get it to generate even something that looks like Degas, for instance. And, I can't even fine tune a painterly art style of any sort into Flux dev. I get that there was working, living artist backlash at SD…

I've had a similar experience, incredible at generating a very specific style of image, but not great at generating anything with a specific style. I suspect we'll see the answer to this is LoRAs. Two examples that stick out are: - Flux Tarot v1 [0] - Flux Amateur Photography [1] Both of these do a great job of combining all the benefits of Flux with custom styles that seem to work quite well. [0] https://huggingface…

I like those, and there's an electroshock lora that's just awesome out there. That said, Tarot and others like it are "illustrator" type styles with extra juice. I have not successfully trained a LoRa for any painting style, Flux does not seem to know about painting.
Post reply on HN