Flux Pro 1.1 is Here! Let's test it.

All Your Tech AI
3 Oct 202411:32

TLDRFlux Pro 1.1, the latest update from Black Force Labs, promises six times faster image generation and improved quality over its predecessor. Codenamed 'blueberry', it outperformed other models in tests, offering better aesthetics and character face details. The update also introduces up to 2K image resolution. Available on Pixel Dojo, users can now access Flux Pro 1.1 to create high-quality images with various prompts, showcasing its capabilities through comparisons with older models.

Takeaways

  • 🚀 Flux Pro 1.1 has been released, a significant update from Black Force Labs, enhancing the capabilities of their pose control tool.
  • 🔍 The new version, previously codenamed 'blueberry', has shown superior performance in mid-journey and Dolly quality and aesthetics.
  • ⏱️ Flux Pro 1.1 offers six times faster generation compared to its predecessor, improving efficiency.
  • 🖼️ The update promises better image quality, prompt adherence, and diversity, with particular improvements to character and human faces.
  • 📊 Flux Pro 1.1 outperformed other models like IDE diagram, mid Journey 6.1, flux Dev, and Dolly 3 in testing.
  • 📈 The new model even surpasses the highly regarded Flux Pro 1.0, which was already considered one of the best models.
  • 📸 An upcoming feature of Flux Pro 1.1 is the ability to generate images up to 2K resolution, a significant leap in image quality.
  • 🛠️ Flux Pro 1.1 is already integrated into Pixel Dojo, allowing users to access the new model through their platform.
  • 📝 The script highlights the importance of prompt crafting for Flux Pro 1.1, as the model's adherence to prompts has changed.
  • 🕷️ A test with a complex prompt about a Coca-Cola bottle with spider features shows Flux Pro 1.1's ability to handle intricate and imaginative requests.
  • 🔧 The video script suggests that users might need to adjust their prompting strategies to get the best results from Flux Pro 1.1 due to changes in the model's text embeddings.

Q & A

  • What is the significance of the announcement of Flux Pro 1.1 by Black Force Labs?

    -The announcement of Flux Pro 1.1 by Black Force Labs is significant because it introduces a new model that provides six times faster generation than the older model, improves image quality, prompt adherence, and diversity, and enhances the quality of human faces, potentially resolving issues like the 'flux chin' seen in previous models.

  • How does Flux Pro 1.1 perform compared to other models in terms of speed and quality?

    -Flux Pro 1.1 outperforms other models like IDE diagram, Mid Journey 6.1, Flux Dev, and even Flux Pro 1.0 in terms of speed and quality. It is noted for its faster generation and improved image quality, making it a significant upgrade.

  • What new feature is expected to be added to Flux Pro 1.1 soon?

    -An upcoming feature for Flux Pro 1.1 is the ability to generate images up to 2K resolution, which will significantly enhance the detail and clarity of the generated images.

  • Where can users currently access and use Flux Pro 1.1?

    -Users can access and use Flux Pro 1.1 on Pixel Dojo. By going to their dashboard and selecting the 'create' tab, they can find various Flux-powered options, including the Flux Image Creator which allows the use of Flux Pro 1.1.

  • What are some of the custom models supported by Flux Dev Laura within Flux Generations?

    -Flux Dev Laura supports up to 20 different custom Laura models directly within the Flux Generations, offering a high level of customization for users.

  • How does the image quality of Flux Pro 1.1 compare to Flux Pro 1.0 when using the same prompts?

    -The image quality of Flux Pro 1.1 appears to be quite different from Flux Pro 1.0, with some images looking more cartoony and CGI-like, while others have more detail and a different aesthetic overall. It suggests that Flux Pro 1.1 may require different prompting techniques to achieve similar results to the previous model.

  • What are the limitations of the Flux Pro 1.1 model in terms of customization?

    -The Flux Pro 1.1 model currently does not support adjustments to the guidance scale or the steps in the API, meaning users are fixed with the default settings internally set by the model.

  • How does the aesthetic appeal of images generated by Flux Pro 1.1 compare to those generated by Flux Pro 1.0?

    -The aesthetic appeal of images generated by Flux Pro 1.1 can be quite different from those of Flux Pro 1.0. While some images show more detail and a unique aesthetic, others may appear less realistic or have a different overall look compared to the original prompts.

  • What changes were made to the large language model in Flux Pro 1.1 that affect prompt adherence?

    -Flux Pro 1.1 saw significant changes to the large language model, or the text embeddings, that transform the prompt into an image. This has resulted in differences in how the model adheres to the prompts, which may require users to adjust their prompting techniques.

  • How does Flux Pro 1.1 handle text-heavy prompts compared to other models?

    -Flux Pro 1.1 is noted for its ability to handle text-heavy prompts well, as demonstrated by the creation of an image of a Coca-Cola bottle with spider features, which came out with a high level of detail and accuracy.

Outlines

00:00

🚀 Introduction to Flux 1.1 Pro and Black Force Labs API

The video script begins with the excitement of a new release from Black Force Labs, the company behind Flux. The script discusses the unveiling of Flux 1.1 Pro and the Black Force Labs API, which is a significant update as it was previously code-named 'blueberry' and had been outperforming other models in quality and aesthetics. The announcement highlights improvements such as six times faster generation, better image quality, prompt adherence, and diversity, particularly in the quality of human faces, potentially resolving issues like the 'flux chin'. The script also mentions impressive test results under the 'blueberry' code name, outperforming other models like IDE diagram, mid Journey 6.1, flux Dev, and Dolly 3. Additionally, the upcoming feature of 1.1 Pro supporting up to 2K image resolution is teased. The speaker has integrated Flux 1.1 Pro into Pixel Dojo, allowing users to create images using various Flux-powered options, including the new 1.1 model, and plans to test it with old Flux Pro prompts to see if there's a visual difference.

05:03

🖼️ Testing Flux Pro 1.1 with Old Prompts

The second paragraph delves into the practical testing of Flux Pro 1.1 by using old prompts to see the visual differences compared to the previous model. The speaker randomly selects images from the Pixel Dojo community gallery, which contains thousands of images generated with Flux Pro, and notes the high quality of the results. The script describes the process of recreating these images with Flux 1.1 Pro, using the same prompts and seeds for consistency. The outcomes are compared side by side, with observations on the differences in aesthetics, realism, and adherence to the prompts. The speaker notes that Flux 1.1 Pro produces images that are less realistic and more cartoony or CGI-like compared to the original Flux Pro. The script also discusses the inability to adjust guidance scale or steps in the API, which are locked to default settings. The speaker saves the new images to a public gallery for viewers to compare and shares their thoughts on the differences in prompt adherence and overall image quality between the models.

10:03

📊 Comparing Flux Pro 1.0 and 1.1 with Various Prompts

The final paragraph of the script continues the comparison between Flux Pro 1.0 and 1.1, using a variety of prompts to illustrate the differences in image generation. The speaker tests Flux 1.1 Pro with prompts that range from a golden tarantula to a woman swimming inside a glass of water, noting the variations in detail and adherence to the original prompts. The script highlights how certain elements, like the anger in a tarantula's expression, are more pronounced in Flux 1.0, while 1.1 Pro produces images that are visually different but still interesting. The speaker also experiments with removing certain prompt descriptors, like 'very angry,' to see how it affects the generated images. The script concludes with a prompt for a Coca-Cola bottle with spider features, demonstrating Flux's ability to handle text-to-image generation effectively. The speaker invites viewers to check out Flux 1.1 Pro on Pixel Dojo and share their thoughts in the comments, signing off as Brian.

Mindmap

Keywords

💡Flux Pro 1.1

Flux Pro 1.1 is the latest version of a pose control tool powered by Flux technology. It is a significant update that promises faster generation times and improved image quality. In the video, Flux Pro 1.1 is compared to its predecessor and other models, showcasing its enhanced capabilities in terms of speed and aesthetics. The script mentions that Flux Pro 1.1 provides six times faster generation than the older model and improves prompt adherence, image quality, and diversity.

💡Black Force Labs

Black Force Labs is the company behind the development of Flux, Flux Pro, and Flux Dev. They announced the release of Flux 1.1 Pro and the Black Force Labs API, which is a big deal for the AI-generated image community. The script indicates that Black Force Labs was testing Flux 1.1 Pro under the code name 'blueberry', which was previously a mystery model that outperformed others in quality and aesthetics.

💡API

API stands for Application Programming Interface, which is a set of rules and protocols for building and interacting with software applications. In the context of the video, the Black Force Labs API allows for the integration of Flux 1.1 Pro into various platforms, enabling users to utilize the new features and capabilities of the updated model.

💡Pixel Dojo

Pixel Dojo is a platform that has integrated Flux Pro 1.1, allowing users to create images using the new model. The video mentions that Pixel Dojo users can access the 'create' tab in their dashboard to use Flux-powered options, including the new Flux Pro 1.1. This integration showcases the practical application of Flux Pro 1.1 in a real-world scenario.

💡Prompt

A prompt in the context of AI-generated images is a text input that guides the model to produce a specific output. The video discusses using prompts to test Flux Pro 1.1 and compares the results with those from older models. Prompts are crucial for directing the AI to create images that match the user's vision.

💡Image Quality

Image quality refers to the visual clarity and detail of an image. The script highlights that Flux Pro 1.1 improves image quality compared to the older model, which is important for users who require high-resolution and detailed images for their projects.

💡Diversity

Diversity in the context of AI-generated images refers to the range of different outputs the model can produce based on the same input. The video mentions that Flux Pro 1.1 improves diversity, meaning it can generate a wider variety of images from the same prompt, offering more creative options to users.

💡Flux Chin

Flux Chin is a term used in the video to describe a common issue with older Flux models where的人物的下巴中间总有一个凹陷。Flux Pro 1.1 is hoped to have resolved this issue, improving the quality of human faces generated by the model.

💡Resolution

Resolution in digital imaging refers to the number of pixels used to form the image, with higher resolutions providing more detail. The script mentions an upcoming feature of Flux Pro 1.1 that will support up to 2K image resolution, which is a significant upgrade from previous models and will allow for even more detailed and high-quality images.

💡Generative AI

Generative AI refers to artificial intelligence systems that can create new content, such as images, based on input data. The video discusses Flux Pro 1.1 as part of a broader suite of generative AI tools available on Pixel Dojo, emphasizing the growing capabilities and applications of this technology.

💡Custom Laura Models

Custom Laura Models are user-specific configurations or presets that can be applied within the Flux Generations to create images. The video mentions that Flux Dev Laura supports up to 20 different custom Laura models, allowing for a high level of personalization and control over the AI-generated images.

Highlights

Flux Pro 1.1 has been released by Black Force Labs, offering significant improvements over previous models.

The new version is six times faster in image generation compared to the older model.

Flux Pro 1.1 enhances image quality, prompt adherence, and diversity.

The model has shown better quality in human faces, potentially resolving the 'flux chin' issue.

Under the code name 'blueberry', Flux Pro 1.1 outperformed other models like Mid Journey and Dolly in quality and aesthetics.

Flux Pro 1.1 is already integrated into Pixel Dojo, allowing users to access the new model.

Users can select Flux Pro 1.1 from the model dropdown in Pixel Dojo's create tab.

Flux Pro 1.1 supports up to 2K image resolution, which is set to be released soon.

Pixel Dojo's community gallery contains thousands of images generated with Flux Pro, showcasing the model's capabilities.

The new model does not currently support steps or guidance adjustments in the API.

Comparisons between Flux Pro 1.0 and 1.1 show differences in visual style and prompt adherence.

Flux Pro 1.1 generated images have a more cell-shaded, cartoony look compared to the more realistic 1.0 version.

The model's performance on text-to-image tasks is impressive, as demonstrated by the creation of a Coca-Cola bottle with spider features.

Flux Pro 1.1's large language model has undergone significant changes, affecting how prompts are transformed into images.

The new model's adherence to prompts is different, requiring experimentation to optimize prompt crafting.

Pixel Dojo offers over 20 generative AI tools, including Flux Pro 1.1, available through a single subscription.