--- license: other license_name: flux1dev tags: - text-to-image - character - comic book - art - graphic novel - flux - flux-diffusers base_model: black-forest-labs/FLUX.1-dev instance_prompt: None widget: - example_title: "Temple" text: Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He navigates through an ancient temple, carefully avoiding the booby traps set in the stone walls. His muscles are taut as he reaches for an artifact glowing on a pedestal, his face showing a mix of caution and determination. Dust fills the air as he steps closer, high-quality graphic art. output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00026_.png - example_title: "Super Wonderman" text: A action scene of graphic novel art character Wonderman wearing a black mask charging himself with super powers, Wonderman a muscular man in a green and red costume with a 'W' emblem on his chest is screaming in pain while being charged by electricity to gain super powers. He is standing with arms wide open consuming the energy around him. high quality graphic art output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00021_.png - example_title: "In Tundra" text: photo realistic still of Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He stands on a frozen tundra, a blizzard raging around him. His body is covered in frost, but he shows no signs of slowing down as he pushes forward through the snow, his eyes focused on a distant mountain peak where an ancient power is hidden. photo output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00036_.png - example_title: "On Moon" text: Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He is walking away from aliens on the moon, high-quality graphic art. output: url: Wonderman Generation Samples/Wonderman1-no-workflow.jpg - example_title: "Fighting" text: A photo realistic comic book character Wonderman wearing a black mask fighting a villain. In the foreground, Wonderman a muscular man in a green and red costume with a 'W' emblem on his chest. The background depicts an action scene of all sorts of fighting characters and a dark, cloudy sky. This is a cinematic action scene that is photorealistic similar to cosplay. photograph output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00023_.png - text: Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He sits in a quiet diner late at night. His mask is still on, but his posture is relaxed as he sips a cup of coffee, watching the rain fall outside. The city is peaceful for now, but Wonderman knows this calm won't last. Modern realistic art style with detailed shading and highlights and high contrast and vivid colors example_title: "w/ Coffee" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00045_.png - text: Wonderman in a black mask a muscular man in a green and red costume with a 'W' emblem on his chest—leaps from a crumbling skyscraper, dodging falling debris while holding a glowing energy sphere in his hand. His black mask is torn, but his face shows fierce determination as he hurls the sphere at an oncoming enemy ship. The sky is filled with smoke and fire from the battle, high-quality graphic art. example_title: "Falling" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00070_.png - text: photo realistic still of Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He stands on a frozen tundra, a blizzard raging around him. His body is covered in frost, but he shows no signs of slowing down as he pushes forward through the snow, his eyes focused on a distant mountain peak where an ancient power is hidden. photo example_title: "In Tundra 2" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00037_.png - text: Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He battles a pack of mutant wolves in an abandoned warehouse, his powerful strikes knocking them back one by one. The moonlight filters through broken windows, casting long shadows as Wonderman moves swiftly, his every motion precise and controlled, high-quality graphic art. example_title: "Wolves" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00030_.png - text: Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He crouches in a rain-soaked alley, muscles tense as thunder rumbles in the background. He grips his glowing energy staff, ready to confront a shadowy figure in the distance. The city lights flicker behind him, high-quality graphic art. example_title: "Kneeling" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00020_.png - text: photo of Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. He stands proudly in front of a massive explosion, framed in the golden hour's soft, warm lighting. His costume is brilliantly contrasted against the fiery background, with the photo perfectly timed to capture the intensity of the scene, high-resolution photograph. example_title: "Explosion" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00046_.png - text: Wonderman wearing a black mask is a muscular man in a green and red costume with a 'W' emblem on his chest. In an action shot, Wonderman speeds through a crowded street, his figure tack sharp while the background is blurred with motion, capturing the sense of speed. The natural light creates a subtle lens flare on his mask in modern realistic art style with detailed shading and highlights and high contrast and vivid colors example_title: "Running" output: url: Wonderman Generation Samples/Wonderman-RunDiffusion-Flux-LoRA_00051_.png license_link: https://huggingface.co/black-forest-labs/FLUX.1-dev/blob/main/LICENSE.md ---
Left Image

Flux Training Concept - Wonderman POC
Darin Holbrook - Chief Technology Officer
RunDiffusion.com / contact@rundiffusion.com

Right Image
# Wonderman Proof of Concept - By RunDiffusion.com ## For this POC we needed to achieve these goals - The concept can not exist in the Flux dataset. (This is cheating) - The concept needed to be present but still allow flexibility for creativity. - The concept needed to resemble the subject within 90% accuracy. - The subject could not "take over" the model. - We used the lowest quality data we could find. (This was easy!) **We chose Wonderman from 1947!** Wonderman is in the public domain, so it can be freely shared, except where restricted by Flux's non-commercial license. Flux thinks that "Wonderman" is "Superman" ![Flux thinks that "Wonderman" is "Superman"](Huggingface-assets/superman-flux.jpg) ## Data Used for Training You can view the [RAW low quality data here: ](https://huggingface.co/RunDiffusion/Wonderman-Flux-POC/tree/main/Raw%20Low%20Quality%20Data). The training data was low resolution, cropped, oddly shaped, pixelated, and overall the worst possible data we've come across. That didn't stop us! AI to the rescue! ![Low Quality Training Data](Huggingface-assets/multiple-samples-training-data.png) To fix the data we had to: - Inpaint problem areas like backgrounds, signatures, and text - Outpaint to expand images - Upscale to get above 1024x1024 at a minimum - Create variations to increase the dataset and provide diverse data We were able to get the dataset to 13 with these techniques. Full dataset [is here](https://huggingface.co/RunDiffusion/Wonderman-Flux-POC/tree/main/Cleaned%20and%20Captioned%20Data) ![Cleaned Wonderman Dataset](Huggingface-assets/multiple-samples-of-cleaned-data.png) ### Captioning the Data We are not entirely familiar with Flux's preferred captioning style. We understand that this model responds will to full descriptive sentences so we went with that. Below are some examples of the images with their captions. We chose LLaMA v3 inspired by this paper: https://arxiv.org/html/2406.08478v1 The system prompt used was basic and could likely benefit from further refinement. A vintage comic book cover of Wonderman. On the cover, there are three main characters: Wonderman in a green costume with a large 'W' on his chest, a woman in a yellow and black outfit, and a smaller figure in a brown costume. Wonderman and the woman appear to be in a dynamic pose, suggesting action or combat. Wonderman is holding a thin, sharp object, possibly a weapon. The woman has a confident expression and is looking towards the viewer. The background is a mix of green and yellow, with some abstract designs. ![Vintage Wonderman](Cleaned and Captioned Data/00008.png) Wonderman, a male superhero character. He is wearing a green and red costume with a large 'W' emblem on the chest. Wonderman has a muscular physique, brown hair, and is wearing a black mask covering his eyes. He stands confidently with his hands by his sides. photo ![Standing Wonderman](https://huggingface.co/RunDiffusion/Wonderman-Flux-POC/resolve/main/Cleaned%20and%20Captioned%20Data/00002.png) ### Train the Data All tasks were performed on a local workstation equipped with an RTX 4090, i7 processor, and 64GB RAM. Note that 32GB RAM will not suffice, as you may encounter out-of-memory (OOM) errors when caching latents. We did use RunDiffusion.com for testing the LoRAs created, enabling us to launch five servers with five checkpoints to determine the best one that converged We're not going to dive into the rank and learning rate and stuff because this really depends on your goals and what you're trying to accomplish. But the rules below are good ones to follow. - We used Ostris's ai-toolkit available here: https://github.com/ostris/ai-toolkit/tree/main - Default config with LR: 4e-4 at Rank 16 - 2200 - 2600 steps saw good convergence. Even some checkpoints into the 4k step range turned out pretty good. If targeting finer details, you may want to adjust the rank up to 32 and lower the learning rate. You will also need to run more steps if you do this. **Training a style:** Using simple captions with clear examples to maintain a coherent style is crucial. Although caption-less LoRAs can sometimes work for styles, this was not within the scope of our goals, so we cannot provide specific insights. **Training a concept:** You can choose either descriptive captions to avoid interfering with existing tokens or general captions that might interfere, depending on your intention. This choice should be intentional. Captioning has never been more critical. Flux "gives you what you ask for" - and that's a good thing. You can train a LoRA on a single cartoon concept and still generate photo realistic people. You can even caption a cartoon in the foreground and a realistic scene in the background! This capability is BY DESIGN - so do not resist it - embrace it! (Spoiler alert next!) ![prompt different backgrounds]() You'll see in the next page of examples where the captioning really helps or hurts you. Depending on your goals again you will need to choose the path that fits what you're trying to accomplish. Total time for the LoRA was about 2 to 2.5 hours. $1 to $2 on RunPod, Vast, or local electricity will be even cheaper. Now for the results! (This next file is big to preserve the quality) ## 500 Steps ![500 steps](Huggingface-assets/500-steps.jpg) ### Direct Use [More Information Needed] ### Downstream Use [optional] [More Information Needed] ### Out-of-Scope Use [More Information Needed] ## Bias, Risks, and Limitations [More Information Needed] ### Recommendations Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations. ## How to Get Started with the Model Use the code below to get started with the model. [More Information Needed] ## Training Details ### Training Data [More Information Needed] ### Training Procedure #### Preprocessing [optional] [More Information Needed] #### Training Hyperparameters - **Training regime:** [More Information Needed] #### Speeds, Sizes, Times [optional] [More Information Needed] ## Evaluation ### Testing Data, Factors & Metrics #### Testing Data [More Information Needed] #### Factors [More Information Needed] #### Metrics [More Information Needed] ### Results [More Information Needed] #### Summary ## Model Examination [optional] [More Information Needed] ## Environmental Impact Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700). - **Hardware Type:** [More Information Needed] - **Hours used:** [More Information Needed] - **Cloud Provider:** [More Information Needed] - **Compute Region:** [More Information Needed] - **Carbon Emitted:** [More Information Needed] ## Technical Specifications [optional] ### Model Architecture and Objective [More Information Needed] ### Compute Infrastructure [More Information Needed] #### Hardware [More Information Needed] #### Software [More Information Needed] ## Citation [optional] **BibTeX:** [More Information Needed] **APA:** [More Information Needed] ## Glossary [optional] [More Information Needed] ## More Information [optional] [More Information Needed] ## Model Card Authors [optional] [More Information Needed] ## Model Card Contact [More Information Needed] - **Developed by:** Darin Holbrook - RunDiffusion co-founder and Chief Technology Officer - **Funded by:** RunDiffusion.com / RunPod.io - **Model type:** Flux [dev] LoRA - **License:** flux1dev https://huggingface.co/black-forest-labs/FLUX.1-dev - **Finetuned from model:** https://huggingface.co/black-forest-labs/FLUX.1-dev