The latest AI image model Flux1.1 is all over the internet! Add the SLR camera file name to get a hyper-realistic image. Netizens: I can't tell the difference

The west wind of the dream morning comes from Aofei Temple
Quantum Bit | Public Account QbitAI

The latest AI literary image model Flux1.1 swept the screen overnight.

With just one simple trick, you can remove the "AI flavor" in the image and achieve photo-quality effects for both people and landscapes.

Reactions from netizens in the comment section were like: I can’t tell the difference, I really can’t tell the difference.

This technique is also very simple to use. Just imitate the file naming format of the SLR camera in the prompt .

For example, "CR2" is the original image file format used by Canon cameras. By entering "IMG " + random number + ".CR2" and then adding the specified content, you can get a realistic image.

Later, netizens who have tried it also reported that they could get good results by switching to Sony camera's "ARW" , Nikon camera's "NEF" , or even Apple's "HEIC" format.

So much so that some people began to doubt whether the model was randomly spitting out a real photo from the training data?

However, if you zoom in on some specific details, it is easy to see that it is indeed AI-generated, such as the garbled text on the license plate.

So is the Flux1.1 model itself very powerful? To what extent does this technique play a role in it?

A senior photo retoucher posted a comparison, with IMG_1018.CR2 added to the left and compared with the right without it. He thought the difference was huge.

Our actual test results also show that adding this technique can significantly improve the realism of the picture.

If you want to try out the Flux1.1 model for free, you can come to the together.ai platform and get $5 in credits when you register.

A random selfie of a tourist at the Great Wall looks authentic at first glance, but a closer look at the texture of the person's skin, the mountains in the background, and the plants still has a whiff of AI.

If we change it to “IMG_0314.cr2: selfie on The Great Wall”, doesn’t it immediately look different?

Codenamed Blueberry, the latest SOTA model

With the official release of FLUX1.1, the mystery of the two unclaimed "blueberry" models that had previously topped the Vincent graph model rankings was also unveiled. It was this one.

The official no longer hides the data and directly releases it. On the Artificial Analysis image arena, FLUX1.1 [pro] , codenamed "blueberry", surpasses all other models and obtains the highest overall Elo score.

In comparison, FLUX1.1 [pro] is cheaper, faster , and its indicators surpass Midjourney, SD3, Ideogram, etc.

In terms of generation speed, FLUX1.1 [pro] is 6 times faster than the previous generation FLUX.1 [pro] while maintaining image quality, command response and improved diversity.

By the way, FLUX.1 [pro] has also been updated and is twice as fast as before. FLUX1.1 [pro] is three times faster than the currently available FLUX.1 [pro].

In addition, officials said that they will soon launch fast high-resolution generation, which FLUX1.1 [pro] can natively support, and can generate 2k images without sacrificing any command response.

FLUX1.1 [pro] will be available through online platforms such as Together.ai, Replicate, fal.ai, Freepik, etc.

At the same time, the official also launched the BFL API, which can be integrated into other developers' own applications. The API pricing is:

• FLUX.1 [dev]: 2.5 cents per image (about RMB 0.18)

• FLUX.1 [pro]: 5 cents per image (about RMB 0.35)

• FLUX1.1 [pro]: 4 cents per image (about RMB 0.28)

Created by the original team of Stable Diffusion

Behind FLUX1.1 [pro] is the original Stable Diffusion team , whose members include Robin Rombach, Andreas Blattmann, Dominik Lorenz, etc.

△ Robin Rombach

In fact, Stable Diffusion was originally an academic research project.

It was led by Professor Björn Ommer, and was completed by Robin Rombach, Andreas Blattmann, Dominik Lorenz and other members of the Machine Vision and Learning Research Group at the University of Munich, as well as Patrick Esser, a researcher at Runway.

Seven months after the research paper was published, Stability AI stepped in to provide computing resources to further develop the text-to-image generation model. In 2022, several of the authors of the above paper joined Stability AI.

Together, the team created Stable Diffusion XL, Stable Video Diffusion, and more.

Rectified Flow Transformers, one of the best papers of ICML 2024 and the Stable Diffusion 3 technical paper, and the Adversarial Diffusion Distillation method used by SDXL-Turbo were also studied by this group of people.

In March this year, it was revealed that these core research team members had resigned collectively.

They then formed a new team called Black Forest Labs , headquartered in Germany.

It was just announced in early August this year and released its first generation of Wenshengtu model FLUX.1. FLUX.1 has three variants: FLUX.1 [pro], FLUX.1 [dev] and FLUX.1 [schnell], balancing performance and accessibility.

Black Forest Labs has completed its seed round of financing, raising a total of $31 million , led by Andreessen Horowitz, followed by Brendan Iribe, Michael Ovitz, Garry Tan, Timo Aila and Vladlen Koltun.

It is said that they have also received follow-up investments from General Catalyst and MätchVC.

Black Forest Labs has also collaborated with Musk to introduce its image generation model into xAI's Grok assistant.

Next, the team revealed that it will launch a SOTA-level text-to-video generation model .

They are said to be raising $100 million at a $1 billion valuation , a significant increase from their previous $150 million valuation.

From Pika 1.5 to Meta Movie Gen, the video generation track has become very popular in the second half of this year. The addition of Black Forest Lab may bring different sparks.

Flux1.1 trial
https://api.together.ai/playground/image/black-forest-labs/FLUX.1.1-pro

Reference links:
[1] https://x.com/fofrAI/status/1841854401717403944
[2] https://blackforestlabs.ai/announcing-flux-1-1-pro-and-the-bfl-api/
[3] https://techcrunch.com/2024/10/03/black-forest-labs-the-startup-behind-groks-image-generator-releases-an-api/

Click here ???? Follow me, remember to mark the star~