PromptZone - Leading AI Community for Prompt Engineering and AI Enthusiasts: Thomas

Creating free images with Flux [pro]

Thomas — Fri, 23 Aug 2024 19:16:23 +0000

Flux [pro] Free

What is it?

Flux [pro] is the most advanced artificial intelligence model for image generation in the [Flux] suite from Black Forest Labs.

This professional version offers enhanced capabilities for creating very high quality visuals of a level equivalent, if not superior, to private models such as Midjourney or DALL-E.

Using Flux [pro]

At the time of its launch, Flux [pro] was only available via paid access from Black Forest Labs partners. Today, however, it's available free of charge from a new site: fluxpro.art.

Here's how to use this online generator to create stunning, fantastic images with Flux [pro] :

Generate an image on fluxpro.art

Go to the [fluxpro.art] website (http://fluxpro.art) and log in by clicking on Login in the top right-hand corner (secure connections via a Google account).
Then open the generator by clicking on Create in the menu at the top of the page.
Enter a prompt in the text field.

For inspiration, have a look at these Flux prompts shared by French-speaking designers.
Finally, click on the Generate button to start the generation process: Flux [pro] works its magic, and in just a few seconds, 4 magnificent images appear.

Flux [pro] parameters

Fluxpro.art lets you modify several generation parameters to make the most of the power of Flux [pro]:

Aspect Ratio designates the image format and determines its dimensions:

Format	Orientation	Width	Height
1:1	Square	1024px	1024px
16:9	Landscape	1344px	768px
9:16 Portrait	768px	1344px
3:2	Landscape	1216px	832px
2:3 Portrait	832px	1216px
5:4	Landscape	1088px	896px
4:5 Portrait	896px	1088px
4:3	Landscape	1152px	864px
3:4	Portrait	864px	1152px

Safety indicates the level of image moderation
- Safe: Suitable for work. Nothing inappropriate.
- Mid : Revealing clothing, violence or slight bloody elements.
- Mature***: Adult themes and situations, partial nudity, graphic violence or death.
- Explicit: Disturbing sexual or graphic content. Graphic nudity, adult objects or contexts.
Steps indicates the number of iteration steps in image generation. A higher number generally generates a more detailed image - but lengthens the generation time and may have undesirable effects on image quality and rendering.
Guidance corresponds to the degree to which the template respects the text. The higher it is, the more the template will try to respect the prompt and the less creative it will be.
Seed determines the random starting state of the image at the start of the generation process. The same seed with exactly the same parameters will produce the same images.

Explore and be inspired

In addition to the Flux [pro] image generator, the fluxpro.art website also features an image gallery containing all the images generated by its users, with their prompts visible. With over 2 million images to choose from, there's plenty of scope for finding thousands of ideas and discovering the talents of AI.

Clicking on an image brings up its prompt and 5 action buttons:

In order, the buttons are :

Copy prompt copies the prompt, which you can then use as you wish.
External Link opens a page with the image and details - ideal for sharing your creations on social networks.
Report allows you to report content for moderation.
Bookmark bookmarks the image, accessible via the Bookmarks link in the main menu.
Like allows you to like an image.

Above the prompt, you'll also see the designer's name (click on it to see his profile and other images) and a Follow button to follow him on fluxpro.art.

Now it's your turn

Now you know everything you need to know to create free images with Flux [pro] and share them with the Fluxpro.art community - all you have to do is create!

Don't forget to take a look at my profile and follow me... I'll be sharing some cool prompts there regularly.

Realistic Photos with Flux: 2026 Prompt Guide

Thomas — Fri, 16 Aug 2024 11:20:46 +0000

If you spend any time on social networks, you won't have missed these ultra-realistic portraits taken with Flux, which have been multiplying in recent days. The craze for Black Forest Labs' new model is real, given its ability to generate impressively realistic photos.

In this article, we'll look at how you can easily use Flux with a LoRA specialized in photorealism to create strikingly realistic images.

This is an English version of this tutorial. A French and a Spanish version are also availabe on my blog.

Flux RealismLora

While the dev and pro versions of Flux are already very good and capable of creating rather realistic images, it's the arrival of the first LoRA Flux-RealismLora a few days ago that seems to have completely changed the game.

The term LoRA (Low-Rank Adaptation) refers to a technique for refining AI models for specific styles or subjects. A LoRA is a file created using this technique and used as a complement to a model. It enables AI to specialize in certain visual aspects without the need for complete retraining.

Here, the LoRA has been specifically trained to reproduce more faithfully the texture, lighting and rendering of photographic images.

Flux on fal.ai

fal.ai is a generative AI platform offering developers easy access to advanced AI models via API.

It enables image generation, text processing and other AI tasks without complex infrastructure. With its pay-per-use pricing and automatic scalability, fal.ai democratizes the integration of AI into applications, making development faster and more cost-effective.

The platform also offers demo and playground areas for AI models, enabling them to be run via a simple web interface. This is the tool we're going to use.

💸 Is it free?

fal.ai charges per use: you only pay for the images you generate: around $0.035 / image.

Registering on fal.ai grants a free credit equivalent to $2.2 - enough to generate about fifty images.

Step-by-step guide

Prerequisite: Github account

Before you start, you'll need a Github account to access the fal platform. If you don't already have an account, create one easily from the Github registration page.

1. Create a fal.ai account

Go to the [fal.ai] website (http://fal.ai) and click on the *Get started* button. Log in with your Github account.

2. Select the Flux Realism LoRA model

Go to the Flux Realism page (also accessible from the fal.ai homepage)

3. Add a prompt

Enter a description of the desired image in the *Prompt.* field. The text must be in English and describe the subject and its environment in detail:

4. Generate an image

All you have to do now is click on the purple *Run* button to start generating the image. As soon as it's ready, it's displayed in the right-hand frame for you to download.

5. To go further

Clicking on the little *More* button under the *prompt* field displays the additional generation parameters:

Flux-Realism prompts

To get the most out of Flux and its Realism Lora, you need to write your prompts in English and structure them clearly and precisely. Here are 7 points to keep in mind and try to include in your prompts to develop detailed and effective prompts:

Main topic: Describe the physical characteristics and attire of the main subject of your photo.
example: middle-aged man with salt-and-pepper hair and a neatly trimmed beard
Action/Pose: Specify the subject's activity, action and/or posture.
example: sitting at a rustic wooden table in a sunlit café.
Context/Location: Detail the environment around your subject.
example: The café backdrop features exposed brick walls, hanging plants, and a chalkboard menu.
Specific details: Refine your description with distinctive elements.
example: wearing a crisp white shirt and navy blazer / The café backdrop features exposed brick walls, hanging plants, and a chalkboard menu.
Ambiance: Define the overall atmosphere.
example: The overall atmosphere is one of urban sophistication and quiet contemplation.
Theme: Explain the context or main activity.
example: His posture and focused expression convey a sense of calm concentration during his morning routine.
Style: Indicate a visual or artistic reference.

example: in the style of amateur photography.

A few prompts

A photo of a middle-aged man with salt-and-pepper hair and a neatly trimmed beard, wearing a crisp white shirt and navy blazer. He's sitting at a rustic wooden table in a sunlit café, carefully sipping from a cappuccino in small cup. His other hand holds an open newspaper. The café backdrop features exposed brick walls, hanging plants, and a chalkboard menu. His posture and focused expression convey a sense of calm concentration during his morning routine. The overall atmosphere is one of urban sophistication and quiet contemplation in the style of amateur photography

A close-up shot of a passionate female chef in her 40s with curly black hair tucked under a white chef's hat, wearing a crisp white double-breasted jacket. She's intently focused on plating a colorful dish in a sleek, modern kitchen. Her skilled hands delicately position a vibrant red bell pepper slice on the plate. Stainless steel appliances and hanging copper pots gleam in the background. Steam rises from the freshly cooked food, creating a misty effect around her. Her expression shows a mix of concentration and pride. The scene captures the artistry and precision of haute cuisine, styled like a high-end culinary magazine spread with dramatic lighting emphasizing the textures and colors of the food.

A wide-angle view of a young Asian male student in his early 20s, wearing a comfortable grey hoodie and jeans, sitting cross-legged on the floor of a vast university library. He's surrounded by open textbooks and papers, with a laptop balanced on his knees. Towering bookshelves filled with leather-bound volumes stretch into the background. Soft, warm lighting from antique desk lamps creates a cozy atmosphere. His brow is furrowed in concentration as he types, with a half-empty coffee cup beside him. Other students can be seen studying at tables in the distance. The image conveys the intensity of exam preparation and the timeless environment of academic pursuit, captured in the style of a thoughtful documentary photograph.

As we've just seen, using Flux to generate realistic images is an accessible and efficient process. The fal.ai platform facilitates this experience by offering an intuitive interface for testing and exploiting Flux's capabilities.

The use of LoRA Realism to generate more lifelike photos demonstrates the full power of a tool like Flux, which is now positioned as a must-have option in the AI-driven image generation landscape. No doubt we'll be hearing more about this model in the future - including on the pages of this blog, where we'll continue to share information and tutorials to democratize the use of such tools.

FLUX: the new prodigy text-to-image model

Thomas — Sat, 03 Aug 2024 07:23:31 +0000

In a context where generative artificial intelligence is growing exponentially, the text-to-image field never ceases to surprise. After the enthusiastic reception of AuraFlow, it's now Flux's turn to make its grand entrance on the scene of alternatives to Stability AI's Stable Diffusion.

Black Forest Labs, a new start-up specializing in generative AI for the media, has just launched Flux, its suite of text-to-image models featuring an open source approach coupled with impressive performance, making it a serious contender not only for Stable Diffusion 3 but also for other industry leaders such as DALL-E and Midjourney.

FLUX.1

Flux's text-image model, trained on a vast dataset of images and captions, has impressive capabilities for generating photorealistic images from natural language descriptions. The model's in-depth understanding of language enables it to interpret complex messages and produce highly detailed, coherent images.

Technical features

Enhanced image quality: Create stunning images at higher resolutions.
Advanced human anatomy and photorealism: Get highly realistic, anatomically accurate images.
Faster adhesion: Get more accurate and relevant images based on your data.
Exceptional speed: Benefit from the speed and efficiency of Flux Schnell, ideal for high-demand applications.
Available in three variants:
- FLUX.1 [dev]: The basic model, shared with a non-commercial license for the community to build on.
- FLUX.1 [schnell]: A distilled version of the basic model, running up to 10 times faster under OpenSource license.
- FLUX.1 [pro]: A private version available only via the API.

How to use FLUX.1

Online feeds

If you'd like to try out a few generations with FLUX.1, you can test the different versions on the Fal.AI and Replicate platforms:

Fal.AI

Replicate

The last two models are also shared on HuggingFace where they can be downloaded.

Download FLUX.1 [dev]

Download FLUX.1 [schnell]

The Black Forest Labs team has shared a sample code on Github that already allows developers and advanced users to run the models on their own machines.

Flux in ComfyUi

The latest update to ComfyUi should already integrate Flux and enable image generation with [dev] and [schell] models.

Workflows have been shared and are available on the Flux example page on the ComfyUi github.

Business strategy and licensing

Black Forest Labs takes an innovative strategic approach with its FLUX suite, offering options tailored to a variety of needs and usage contexts:

FLUX.1 [pro] : The spearhead of the range, this high-end model promises unrivalled performance. Accessible only via the Black Forest Labs API, it targets professionals demanding exceptional visual quality. Although pricing details have yet to be confirmed, we can expect a premium commercial model.
FLUX.1 [dev] : An intermediate version, FLUX.1 [dev] offers a balance between performance and accessibility. Designed for developers and researchers, it enables fine-tuning and more flexible use, while remaining framed by non-commercial terms of use. For commercial use, specific negotiations with Black Forest Labs will be necessary.
FLUX.1 [schnell]: A true gateway to the FLUX universe, this speed-optimized version sacrifices some quality for accessibility. Distributed under an open source license (Apache 2.0), FLUX.1 [schnell] opens up a vast field of possibilities for large-scale experimentation and innovation.

This three-pronged strategy illustrates Black Forest Labs' determination to democratize access to generative AI while preserving a viable business model. It thus offers a range of solutions to meet the varied needs of professionals, researchers and AI enthusiasts alike.

Future prospects

The launch of FLUX by Black Forest Labs marks a turning point in the text-to-image industry. Backed by a $31 million Seed round led by Andreessen Horowitz, the company is well positioned to influence the future of generative AI.

The Black Forest Labs team, made up of renowned researchers and engineers who have contributed to major innovations such as VQGAN and Stable Diffusion, is already planning the development of text-to-video systems. This approach could accelerate innovation in fields as diverse as cinema, advertising and education.

By democratizing access to these cutting-edge technologies while emphasizing transparency and security, Black Forest Labs aims to shape a more open, collaborative and innovative generative AI ecosystem. Time will tell whether FLUX will be able to hold its own against the giants of the sector, but one thing is certain: competition in the field of text-to-image AI has only just begun.

Stable Diffusion Prompt Weights: 2026 Complete Guide

Thomas — Thu, 01 Aug 2024 13:58:40 +0000

Since its introduction, Stable Diffusion has revolutionized the way we create images: anyone can use AI to create images from a textual description of what they want to represent. But while this works pretty well, you'll have found that knowing how to write that description is essential to getting satisfactory results. That's why we're going to take a look in this article at a simple but effective technique for better controlling image generation with Stable Diffusion: Prompt Weighting.

ℹ️ This article is a translation from the French article Prompt Weight : le poids des mots originally published on my Stable Diffusion Blog.

Prompt Weighting is a tool that allows you to give more or less importance to certain parts of the text you submit to Stable Diffusion. In other words, it's a way of guiding the AI's attention to the key elements you want to appear in the generated image.

It may sound trivial, but in reality, this feature opens the door to very powerful creative control - especially as it doesn't require you to install any other complex templates or plugins: it's just a matter of using a particular syntax in the prompt.

Let's explore in detail what Prompt Weighting is, how it works and, above all, how it can help you take your creativity to the next level.

Prompt Weighting

What is Prompt Weighting?

In practical terms, Prompt Weighting uses the principle of weighting to change the relative importance of concepts or words in your prompt by changing their Weight.

Increasing word weights

Let's imagine a simple (and simplistic) prompt like "Woman, Beach, Pizza". In this prompt, each of the words has the same importance as the others - they all have a Weight of 1. Put another way, each one weighs 1/3 of the prompt's total.

To increase the importance of a word, and influence the result accordingly, we need to increase its weight, so that it weighs more than the others in the prompt total.

Let's compare 4 image series, generated with the same seed but varying word weights:

As you can see, the images remain quite similar, but by changing the prompt weight to give more weight to a word, its influence on the image is more marked.

Look, for example, at how, when importance is given to "Beach", the hills and cliffs in the background disappear to give more room to the beach and the sea. Or how the pizza is a little larger when this word is given more weight.

Every word counts

The relative importance of each word also depends on the length of the prompt. In our simplified example, we only had 3 words - so each one is already very important in the prompt (1/3 of the total).

When there are more words in the prompt, the relative importance of each word decreases. And increasing the weight of any one of them can have a greater effect than in this first example.

Let's take a similar but longer and more detailed prompt for a new test: Photography of a Woman with dark hair and blue eyes eating a slice of pepperoni pizza on a white beach: at sunset, ocean, cinematic shot, natural light

The most noticeable difference is obtained when the weight of "beach" is increased, as this leads to wider shots leaving more room for the beach in the image. This is particularly noticeable in the first two images of the series.

As you can see, Prompt Weighting allows you to refine the description of your image to give more importance to certain words or expressions and thus modify the result.

How do I use it?

Most Stable Diffusion interfaces allow you to vary the weight of words directly in the prompt - the relative importance of each word being calculated before image generation.

Each interface has its own way of implementing this feature - but the way of using it is quite similar from one to another.

Prompt Weight in Automatic1111

Automatic1111 was one of the first interfaces to implement the use of Prompt Weight using a syntax based on () and [] - which inspired the syntax for most modern interfaces.

The simple way to change the weight of a word in Automatic is therefore to enclose the word in () to increase its weight - and enclose it in [] to decrease its weight. For example, "Photography of a woman with (blue) eyes" for a prompt that reinforces the importance of the color blue (blue). And "Photography of a woman with [blue] eyes" to decrease it.

For greater precision, use a slightly different syntax: the word or expression is enclosed in brackets, followed by ":" and the weight modifier you wish to add. For example, "Photography of a man wearing jeans and a (pink:1.5) shirt" will increase the weight of the color pink (pink). To decrease the importance of a word, use a weight less than 1.

In practice, (keyword) is equivalent to (keyword:1.1) and [keyword] is equivalent to (keyword:0.9).

Prompt Weight in Fooocus

Fooocus uses the same syntax as Automatic1111 and applies it in the same way. The aim is to make it easy to copy/paste image prompts created with Automatic1111 to obtain similar results with Fooocus.

The Fooocus interface also implements a shortcut for using this function: by selecting a word or expression, it is possible to change its weight by pressing Ctrl+⬆️ to increase it and Ctrl+⬇️ to decrease it.

Prompt Weight in ComfyUI

The syntax used in ComfyUI to change the weight of a word is also very similar to that of Automatic1111.

ComfyUI uses the () accompanied by the weight, such as (keyword:1.1) to adjust the weight of a word or expression in the prompt.

You can also select a word or expression and press Ctrl+⬆️ and Ctr+⬇️ to increase or decrease its weight. The extent to which you increase or decrease the weight of these shortcuts can be adjusted in the parameters.

Difference between Automatic1111 and ComfyUI

Even though the concept is the same and the syntax is practically the same, Automatic1111 and ComfyUI's prompt weighting is different.

Indeed, the way weights are applied is not the same in the two interfaces. ComfyUI processes and applies weights as specified, whereas Automatic1111 will normalize the weights so that they sum to 1.

In practice, this means that the same weight in ComfyUI will generally have a stronger effect than in Automatic1111 - so it's best to use lower values in ComfyUI than in Automatic1111.

Indeed, with the same prompt weights, A1111 generally has a very weak effect compared to ComfyUI. You therefore need to apply much higher weights in A1111 to achieve a similar effect. And higher weights often produce poorer quality, more "cartoony" results in ComfyUI.

Because of this difference, it's also very difficult to replicate the same results between the two interfaces, even with identical parameters, as the weighting of words in the prompt is not ultimately the same between the two interfaces.

Fooocus has chosen to apply the same normalization calculation as Automatic1111 to enable reuse of prompts from one to the other.

Prompt Weighting is therefore a powerful technique for fine-tuning and precisely controlling the generation of images by Stable Diffusion. By adjusting the weight of words and phrases in your prompts, you can subtly or radically influence the final result, opening up new creative possibilities.

By mastering this technique, you can refine your creations and achieve results closer to your artistic vision. Whether you're an occasional user or a seasoned artist, this technique gives you greater control over the AI creative process.

Feel free to experiment with different weights and combinations to discover the full potential of this feature. Prompt Weighting is an invaluable tool in your creative toolbox, enabling you to push the boundaries of what's possible with Stable Diffusion.

Fooocus 2.5 is here

Thomas — Tue, 30 Jul 2024 18:49:38 +0000

Although Fooocus aims to provide a simple and intuitive interface, it can't ignore the advancements in Stable Diffusion and the new uses of generative AI.The application might have seemed neglected by its creator, but a collaborator, mashb1t, has taken over in recent months and is handling the updates. The recently released version 2.5.0 owes much to mashb1t's efforts—let's take a closer look at what it has to offer.

How to update Fooocus ?

Typically, Fooocus updates automatically when you run it using the run.bat file on Windows or the entry_with_update.py script on Mac. If the automatic update doesn't happen, you'll need to update manually using one of the following commands:

For Windows: ..\python_embeded\python.exe -m pip install -r .\requirements_versions.txt
For Mac: python.exe -m pip install -r requirements_versions.txt

These commands will install the necessary dependencies as specified in the requirements_versions.txt file.

What's New in Fooocus 2.5.0 ?

Enhance: Correct and Improve

The new "Enhance" feature in Fooocus 2.5.0 allows users to enlarge, improve, or correct images or specific parts of them. Similar to ADetailer, Enhance offers greater flexibility and better object detection. It is particularly effective for refining and enhancing faces and intricate details in an image:

Enhance also enables targeted and automatic inpainting to modify specific details, such as changing the color of a dress or correcting hand details:

As with many features in Fooocus, Enhance leverages a combination of specialized models to achieve the best possible results. The primary models used are SAM (Segment Anything) and GroundingDINO, though Fooocus also supports other models.

New Models and Presets

Some presets in Fooocus have been updated to utilize the latest versions of the models:

The Anime preset now uses AnimaPencilXL_v500.
The Realistic preset has been updated to use LoRA SDXL_FILM_PHOTOGRAPHY_STYLE_V1.
A new preset, pony_v6, has been introduced, featuring the popular ponyDiffusionV6XL. This preset includes a "Fooocus Pony" style, which automatically applies essential tags for this model, such as "score_9," "score_8_up," and "score_7_up."

Other Notable Updates

A new checkbox to enable inpainting masks is now available in the Inpainting or Outpainting tab.
Image editing features have been added for uploaded inpainting masks, allowing for more precise control over edits.
The inclusion of VAE (Variational Autoencoder) has been added as a configurable option within presets, giving users more flexibility in adjusting image generation settings.

🔗 For more details, please refer to the release notes for versions 2.5.0 on Fooocus's GitHub page.

ℹ️ This article is a translation from the Fooocus 2.5 announcement on the French Stable Diffusion Blog.