7 ms·
Releasing weights for FLUX.1 Krea
- dvrp 1y agoHello everyone. I’m the Co-founder and CTO of Krea. We’re excited because we wanted to release the weights for our model and share it with the HN community for a long time. My team and I will try to be online and try to answer any questions you may have throughout the day.
- jackphilson 1y agoHi. Thanks for this. What is your goal of doing so? From a business standpoint. Or is it purely altruistic?
- dvrp 1y agoHaha-classic! It’s simple: hackability and recruiting! The open-source community hacking around it and playing with it PLUS talented engineers who may be interested in working with us already makes this release worth it. A single talented distributed systems engineer has a lot of impact here. Also, the company ethos is around AI hackability/controllability, high-bar for talent, and AI for creatives - so this aligns perfectly. The fact that Krea serves both in-house and 3rd-Party models tells you that we are not that bullish on models being a moat.
- wjrb 1y agoI can say that it's definitely working on me! I hadn't heard of Krea before, and this is a great introduction to your work. Thanks for sharing it.
- cchance 1y agoPeople underestimate how much goodwill companies gain from pushing opensource stuff out, not just from word of mouth but even picking up users for their commercial offerings too, while i could run opensource and appreciate it in a lot of cases using API's from the companies that i like (mostly ones that do opensource stuff) tends to be easier for bigger stuff...
- mk_stjames 1y agoAny plans to get into working with the Flux 'Kontext' version, the editing models? I think the use cases of such prompted image editing is just wildly huge. Their demo blew my mind, although I haven't seen the quality of the open weight version yet. It is also a 12B distill.
- cubefox 1y agoRegarding the P(.|photo) vs P(.|minimal) example, how do you actually decide this conflict? It seems to me that photorealism should be a strong default "bias". My reasoning: If the user types in "a cat reading a book" then it seems obvious that the result should look like a real cat which is actually reading a book. So it obviously shouldn't have an "AI style", but it also shouldn't produce something that looks like an illustration or painting or otherwise unrealistic. Without further context, a "cat" is a photorealistic cat, not an illustration or painting or cartoon of a cat. In short, it seems that users who want something other than realism should be expected to mention it in the prompt. Or am I missing some other nuances here?
- Western0 1y agoI need model for other language than english
- dvrp 1y agoFor Dang or HN-mods: I noticed that the URL for this submission is wrong: I tried to submit the correct URL (https://www.krea.ai/blog/flux-krea-open-source-release https://www.krea.ai/blog/flux-krea-open-source-release) but, for some reason, the submission gets flagged as duplicated and then I can only find this item which has a URL to our old blog post. In the mean time, I'll setup a server-side redirect from the old blog post to our new one, but it would be nice to fix the link and I don't think I can do it on my side.
- dang 1y agoIt's because https://www.krea.ai/blog/flux-krea-open-source-release https://www.krea.ai/blog/flux-krea-open-source-release contains this: <link rel="canonical" href="https://www.krea.ai/blog/new-krea"> Our software follows canonical urls when it finds them. I've fixed the link above now (and rolled back the clock on the submission, to make up for lost time) but you might want to fix this for future pages.
- dvrp 1y agoOMG. Thank you! I had to setup a CDN-level redirect and I was so confused as to why when I asked others to help, their submissions were flagged as [dupe] or [dead] Thank you so much! I knew that HN software was advanced, but I didn’t know you guys used Canonical URLs like Google does. Smart and thanks for helping us with this slip!!!
- dang 1y agoOh you're welcome - it does lead to a lot of not-obvious problems like this but I think it's worth it overall. It helps with duplicate detection, merging threads, and so on.
- sangwulee 1y agoHi! I'm lead researcher on Krea-1. FLUX.1 Krea is a 12B rectified flow model distilled from Krea-1, designed to be compatible with FLUX architecture. Happy to answer any technical questions :)
- swyx 1y agothanks for doing this! what does " designed to be compatible with FLUX architecture" mean and why is that important?
- sangwulee 1y agoFLUX.1 is one of the most popular open weights text-to-image models. We distilled Krea-1 to FLUX.1 [dev] model so that the community can adopt it seamlessly into existing ecosystem. Any finetuning code, workflows, etc that was built on top of FLUX.1 [dev] can be reused with our model :)
- oompty 1y agoThe model looks incredible! Regarding this part: > Since flux-dev-raw is a guidance distilled model, we devise a custom loss to finetune the model directly on a classifier-free guided distribution. Could you go more into detail on the specific loss used for this and any other possible tips for finetuning this that you might have? I remember the general open source ai art community had a hard time with finetuning the original distilled flux-dev so I'm very curious about that.
- vipermu 1y agohey hn! I'm one of the founders at Krea. we prepared a blogpost about how we trained FLUX Krea if you're interested in learning more: https://www.krea.ai/blog/flux-krea-open-source-release https://www.krea.ai/blog/flux-krea-open-source-release
- orphea 1y agoOff topic but did you really hide scroll bars on the website? Why...? .scrollbar-hide { -ms-overflow-style: none; scrollbar-width: none; }
- VladVladikoff 1y agoUI brought to you by vibe code
- BoorishBears 1y agonah, Krea is just from that side of design twitter where you don't uppercase letters and you can break the rules sometimes. very atypography-coded.
- johnisgood 1y agoThey probably did it because the website might look better without a scrollbar, but they should realize that many browsers hide the scrollbar and they only get displayed when you hover over or when you start scrolling. That said, the scrollbar is always there for me (unless hidden by CSS), and I would not have minded it at all.
- erwannmillon 1y agoyoo i'm also a researcher on the krea 1 project and happy to answer any questions :)
- SkannR 1y ago[dead]
- gutianpei 1y agohello Erwann, great work! I have a very technical question just for you: how are you today?
- erwannmillon 1y agohahahah i'm doing well tianpei good to hear from you!
- OsrsNeedsf2P 1y agoCool to see an open weight model for this. But what's the business use case? Is it for people who want to put fake faces on their website that don't look AI generated?
- dvrp 1y agoThanks! From a business point of view, there are many use-cases. Here's a list in no particular order: - You can quickly generate assets that can be used _alongside_ more traditional tools such as Adobe Photoshop, After Effects, or Maya/Blender/3ds Max. I've seen people creating diffuse maps for 3D using a mix of diffusion models and manual tweaking with Photoshop. - Because this model is compatible with the FLUX architecture, we've also seen people personalizing the model to keep products or characters consistent across shots. This is useful in e-commerce and fashion industry. We allow easy training in our website — we labeled it Krea 1 — to do this, but the idea with this release is to encourage people with local rigs and more powerful GPUs to be able to tweak with LoRAs themselves too. - Then I've seen fascinating use-cases such as UI/UX designers who prompt the model to create icons, illustrations, and sometimes even whole layouts that then they use as a reference (like Pinterest) to refine their designs on Figma. This reminds me of people who have a raster image and then vectorize it manually using the pen tool in Adobe Illustrator. We also have seen big companies using it for both internal presentations and external ads across marketing teams and big agencies like Publicis. EDIT: Then there's a more speculative use-case that I have in mind: Generating realistic pictures of food. While many restaurants have people who either make illustrations of their menu items and others have photographers, the big tail of restaurants do not have the means/expertise to do this. The idea we have from the company perspective is to make it as easy as snapping a few pictures of all your dishes and being able to turn all your menu (in this case) into a set of professional-looking pictures that accurately represent your menu.
- gshklovski 1y agoHelpful blog post for understanding what kind of data is needed for these models! Does this have any application for generating realistic scenes for robotics training?
- dvrp 1y agoHey thanks! I’ll ask Sangwu to hop here to answer this and give a more research-oriented answer
- sangwulee 1y agoThank you! Glad you find it helpful. The model is focused on photorealism so it should be able to generate most realistic scenes. Although, I think using 3D engines would be more suitable for typical cases for robotics training since it gives you ground truth data on objects, location, etc. One interesting use case would be if you are focusing on a robotics task that would require perception of realistic scenes.
- erwannmillon 1y agoyeah data really is everything that was the number one lesson from this whole project
- kmavm 1y agoAmazing. I can practically smell that owl it looks so darned owl-like. From the article it doesn’t seem as though photorealism per se was a goal in training; was that just emergent from human preferences, or did it take some specific dataset construction mojo?
- sangwulee 1y agoI love owls. Photorealism was one of the focus areas for training because "AI look" (e.g. plastic skin) was biggest complaint for FLUX.1 model series. Photorealism was achieved with both careful curation of finetuning and preference dataset.
- bangaladore 1y agoCan someone ELI5 why the safetensor file is 23.8 GB, given the 12B parameter model? Does the model use closer to 24 GB of VRAM or 12 GB of VRAM. I've always associated a 1 billion parameter = 1 GB of VRAM. Is this estimate inaccurate?
- petercooper 1y agoThat's a good ballpark for something quantized to 8 bits per parameter. But you can 2x/4x that for 16 and 32 bit.
- 7734128 1y agoI've never seen a 32 bit model. There's bound to be a few of them, but it's hardly a normal precision.
- zamadatix 1y agoSome of the most famous models were distributed as F32, e.g. GPT-2. As things have shifted more towards mass consumption of model weights it's become less and less common to see.
- petercooper 1y agoAnd on the topic of image generation models, I think all the Stable Diffusion 1.x models were distributed in f32.
- nodja 1y ago> As things have shifted more towards mass consumption of model weights it's become less and less common to see. Not the real reason. The real reason is that training has moved to FP/BF16 over the years as NVIDIA made that more efficient in their hardware, the same reason you're starting to see some models being released in 8bit formats (deepseek). Of course people can always quantize the weights to smaller sizes, but the master versions of the weights is usually 16bit.
- 1y ago
- dvrp 1y agoRelevant links: - GitHub repository: https://github.com/krea-ai/flux-krea https://github.com/krea-ai/flux-krea - Model Technical Report: https://www.krea.ai/blog/flux-krea-open-source-release https://www.krea.ai/blog/flux-krea-open-source-release - Huggingface model card: https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev
- TuringNYC 1y agoI usually use https://github.com/axolotl-ai-cloud/axolotl https://github.com/axolotl-ai-cloud/axolotl on Lambda/Together for working with these types of models. Curious what others are using? What is the quickest way to get started? They mention Pre-training and Post-training but sadly didnt provide any reference starter scripts.
- dvrp 1y agoWe actually have a GitHub repository to help with inference code. Check this out: https://github.com/krea-ai/flux-krea https://github.com/krea-ai/flux-krea Let me see if we can add more details on the blog post and thanks for the flag!
- TuringNYC 1y agoThanks! Yes, the inference is pretty straightforward, but the real opportunity IMHO is the custom pre-training and post-training opportunities given the open weights.
- sergiotapia 1y agoFor the Krea team that might be reading: I was trying to evaluate Krea for my image gen use case, and couldn't find: - cost per image - latency per image Hope you guys can add it somewhere!
- dvrp 1y agoYup... We have the following: https://www.krea.ai/pricing https://www.krea.ai/pricing Though we wanted to keep this technical blogpost free from marketing fluff, but maybe we over-did it. However, sometimes it's hard to give an exact price per image, as it depends on resolution, number of steps, whether a LoRA is being used or not, etc.
- bluehark 1y agoDo you have an NVIDIA optimized version? Similar to how RTX accelerated FLUX.1 Kontext: https://blogs.nvidia.com/blog/rtx-ai-garage-flux-kontext-nim-tensorrt/ https://blogs.nvidia.com/blog/rtx-ai-garage-flux-kontext-nim...
- sangwulee 1y agoWe have not added a separate RTX accelerated version for FLUX.1 Krea, but the model is fully compatible with existing FLUX.1 dev codebase. I don't think we made a separate onnx export for it though. Doing 4~8 bit quantized version with SVDQuant would be a nice follow up so that the checkpoint is more friendly for consumer grade hardware.
- bluehark 1y agoHow large was the dataset used for post-training?
- sangwulee 1y agoWe used two types of datasets for post-training. Supervised finetuning data and preference data used for RLHF stage. You can actually use less than < 1M samples to significantly boost the aesthetics. Quality matters A LOT. Quantity helps with generalisation and stability of the checkpoints though.
- lawlessone 1y agoHow is the data collected?
- sangwulee 1y agoThe highest quality finetuning data was hand curated internally. I would say our post training pipeline is quite similar to SeedDream 2.0 ~ 3.0 series from ByteDance. Similar to them, we use extensive quality filters and internal models to get the highest quality possible. Even from there, we still hand curate a hand-picked subset.
- qotgalaxy 1y ago[dead]
- qlm 1y agoImages still look AI generated
- dvrp 1y agoYeah, there are still imperfections. But, it’s surprising to us how much the quality can be improved without the need of a whole pre-training (re-) run.
- Lerc 1y agoIs it possible (or do people already do this), to train a classifier to identify the AI look and use it as an adversary to try and maximise both 'quality' and 'not that sort of quality'?
- sangwulee 1y agoI actually tried a few experiments in early exploration stages! I trained a small classifier to judge AI vs non-AI images. Use it as a reward model to do small RL / post training experiments. Sadly, was not too successful. We found that directly finetuning the model on high quality photorealistic image was most reliable. Another note about preference optimisation and RL is that it has really high quality ceiling but needs to be very carefully tuned. It's easy to get perfect anatomy and structure if you decide to completely "collapse" the model. For instance, ChatGPT images are collapsed to have slight yellow color palette. FLUX images always have this glossy, plastic texture with overly blurry background. It's similar to reward hacking behavior you see in LLMs where they sound overly nice and chatty. I had to make a few compromises to balance between "stable, collapsed, boring model" and "unstable, diverse, explorative" model.
- Lerc 1y agoI could see how you might need a multi channel classifier so that one exists on a range (A) of -1 = "This looks like AI" to 1="This does not look like AI" and another(R) where 1="The above factor is relevant to this image" to 0="The AI-ness of this image is not a meaningful concept Then optimise for max (Quality + A*R) Arguably amplitude of A should do R but I think the AI-ness and the AI-ness-relevance are distinct concepts (It could be highly relevant but it can't tell what it should be).
- artninja1988 1y agoThanks for the release to the team! Any plans on doing the same for FLUX.1 Kontext?
- leftstrokeviral 1y agoHow much data is the model trained on?
- dvrp 1y agoCopying and pasting Sangwu’s answer: We used two types of datasets for post-training. Supervised finetuning data and preference data used for RLHF stage. You can actually use less than < 1M samples to significantly boost the aesthetics. Quality matters A LOT. Quantity helps with generalisation and stability of the checkpoints though.
- lawlessone 1y agoHow is data acquired and curated?
- paparicio 1y agoGreat, these people are crazy. CONGRATS!!!
- dvrp 1y agoHaha thanks! Do you happen to have a use-case for it?
- TuringNYC 1y agoI'd recommend you offer a clearly documented pathway for companies to license commercial output usage rights if they get the results they seek (i'll know soon enough!)
- dvrp 1y agoYou can find details about the license here: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/mai... In a nutshell, it follows the same license as BFL Flux-dev model.
- ilc 1y agoTried a simple prompt, and got some pretty interesting results: "Octopus DJ spinning the turntables at a rave." The human like hands the DJ sprouts are interesting, and no amount of prompting seems to stop them. Opinionated, as the paper says.
- earthicus 1y agoDescribing it as "Octopus DJ with no fingers" got rid of the hands for me, but interestingly, also removed every anthropomorphized element of the octopus, so that it was literally just an octopus spinning turntables.
- ilc 1y agoI still get octopus hands, even with just "Octopus DJ with no fingers." nothing else. Maybe you got a lucky roll :)
- SkannR 1y ago[dead]
- jacooper 1y agoNon-commercial license... What's the point even, I can't use it for anything.
- hhh 1y agofor the love of the game, not for corporate greed
- codedokode 1y agoThe license, as I understand, applies only to the model, not to the images produced? Otherwise they should respect the license of images it was trained on.
- lawlessone 1y agoHow did you train while ensuring only images consensually acquired were used?
- whimsicalism 1y agoLikely the same way visual artists ensure that they only learn from images with permissive licenses.
- codedokode 1y agoHuman learning and computer processing millions of works are different things. I don't think any human artists have seen as many images as the developers used for training.
- renewiltord 1y agoThen let he who hath not sinned cast the first stone.
- jamespo 1y agoYes scale is totally irrelevant, that's what all of FAANG tell us too :/
- renewiltord 1y agoGuy who gets his morality from big company internal policies.
- 0x457 1y agoOnly if you don't count artists with able eye(s) living their life?
- throw10920 1y agoLooking at real things while living life is categorically different from viewing millions of artworks made by other artists. I can guarantee you that no artist has ever seen millions of artworks. This comment is not even bad-faith, it's just wrong.
- CaptainFever 1y agoNitpick: this is not open weights, this is weights available. The license restricts many things like commercial, NSFW, etc.
- dragonwriter 1y agoI mean this started with Stable Diffusion 1.x->XL which were only loosely open, and has just gotten worse with progressively farther from open licensed image gen models being described as “open weights”, but, yes, Flux.1 Krea (like the weights-available versions of Flux.1 from BFL itself) is not open even to the degree of the older versions of Stable Diffusion; weights available and “free-as-in-beer licensed for certain uses”, sure, but not open.
- dang 1y agoAlright we've made the title not say open, in the hope of routing around this objection.
- SubiculumCode 1y agoI've never gotten one to make what I am thinking of: A Galton board. At the top, several inches apart are two holes from which balls drop. One drops blue balls, the other red balls. They form a merged distribution below in columns, demonstrating dual overlapping normal distributions Imagine one of these: https://imgur.com/a/DiAOTzJ https://imgur.com/a/DiAOTzJ but with two spouts at the top dropping different colored balls Its attempts: https://imgur.com/undefined https://imgur.com/undefined https://imgur.com/a/uecXDzI https://imgur.com/a/uecXDzI
- CGMthrowaway 1y agoHave you tried building one irl? I can't find a video of a double one
- SubiculumCode 1y agoI have not. It definitely is not something in training sets :)
- vunderba 1y agoNice release. Ran some preliminary tests using the 12b Txt2Img Krea model. Its biggest wins seems to be raw speed (and possibly realism) but perhaps unsurprisingly did not score any higher on the leaderboard for prompt adherence than the normal Flux.1D model. https://genai-showdown.specr.net https://genai-showdown.specr.net On another note, there seem to be some indication that Wan 2.2+ future models might end up becoming significant players in the T2I space though you'll probably need a metric ton of LoRAs to cover some of the lack of image diversity.
- dvrp 1y agoCan you point to a URL with the tests you’ve done? Also, FWIW, this model focus was around aesthetics rather than strict prompt adherence. Not to excuse the bad samples, but to emphasize what was one of the research goals. It’s a thorny trade-off, but an important one if one wants to get rid of what’s sometimes known as “the flux look”. Re: Wan 2.2 I’ve also been reading of people commenting about using Wan 2.2 for base generation and Krea for the refiner pass which I thought was interesting.
- vunderba 1y agoThe Image Showdown site actually does have Flux Krea images but they're hidden by default. If you open up the "Customize Models" dialog you can compare them against other Flux models (Flux.1 Dev and Kontext). > FWIW, this model focus was around aesthetics Agreed - whereas these tests are really focused on various GenAI image models ability to follow complicated prompts and are not as concerned with overall visual fidelity. Regarding the "flux look" I'd be interested to see if Krea addresses both the waxy skin look AND the omnipresent shallow depth of field.
- littlestymaar 1y agoI saw this comparison on reddit[1] between Wan2.2 and FLUX.1 Krea and it doesn't look like Krea is successful at avoiding the “AI look”, whereas Wan succeed brilliantly. [1]: https://www.reddit.com/r/StableDiffusion/comments/1mec2dw/texttoimage_comparison_flux1_krea_dev_vs/ https://www.reddit.com/r/StableDiffusion/comments/1mec2dw/te...
- buyucu 1y agoDoes it work out of the box with other Flux-compatible tools such as sd-scripts?
- Western0 1y agouv not working no clicking no torch and Cannot access gated repo for url https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev/resolve/main/ae.safetensors https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev/res.... Access to model black-forest-labs/FLUX.1-Krea-dev is restricted. You must have access to it and be authenticated to access it. Please log in.
- shreyank06 1y agoI recently ran a training experiment using the same dataset, number of steps, and epochs on both Flux Dev and Flux Krea models. What stood out to me was that Flux Dev followed the text prompts more accurately, whereas Krea’s generations were more loosely aligned or "off" in terms of prompt fidelity with deformations in body type and the architecture. Does this suggest that Flux Krea requires more training to achieve strong text-to-image alignment compared to Flux Dev? Or is it possible that Krea is optimized differently (e.g. for style, detail, or artistic variation rather than strict prompt adherence)? Curious if anyone else has experienced this or has any insight into the differences between these two. Would love to hear your thoughts