BRDRLNDS - Pandora's Style Unleashed!

Bring the iconic, graphic-novel-inspired look of the Borderlands to your generations!

---

This LoRA, "BRDRLNDS", is inspired by the distinctive art style of the Borderlands game series (1, 2, Pre-Sequel, 3, Wonderlands, and 4). It aims to replicate the unique blend of hand-drawn textures, bold outlines, and vibrant, sometimes gritty, environments and characters that define the franchise.

Whether you're aiming for a Vault Hunter portrait, a bizarre alien landscape, or a scene of chaotic combat, this LoRA can help infuse your creations with that unmistakable Pandora flair. The style is often described as a "graphic novel" or "concept art" approach rather than traditional cel-shading, characterized by thick black lines and hand-painted textures.

Trained specifically for the Blackforest Labs Flux.1 Dev model, this LoRA is designed to leverage Flux's capabilities for detail and composition while applying the stylized aesthetic.

---

No trigger words required!

This LoRA was trained using detailed captions describing the image content alongside the style. This approach allows the LoRA to apply the Borderlands style more flexibly to a wider range of subjects without needing a specific trigger keyword. Simply include descriptive terms in your prompt, and the LoRA, when active, should influence the output towards the desired style.

---

Recommended Settings

• LoRA Weight: 0.8 - 1.0 (Start here and adjust based on desired intensity)

• Base Model: Blackforest Labs Flux.1 Dev

• Sampler: Euler Beta

• Steps: 30-35+

• CFG Scale: 1

• Resolution: 1024x1024 or relevant aspect ratios like 1024x1536, 1536x1024, 1920x1088, 1344x768, etc. (Flux.1 Dev supports various resolutions up to 2MP, but 1024x1024 is a good starting point).

---

• Excels at applying the Borderlands aesthetic to subjects present in the training data, such as:

  • Human characters (Vault Hunters, bandits, etc.)

  • Alien creatures and animals

  • Post-apocalyptic/sci-fi environments (deserts, industrial areas, alien worlds)

  • Vehicles and weaponry common in the series.

• Capable of generating a variety of subject matter within the style.

---

• May struggle to apply the style as effectively to subjects drastically different from the training dataset.

• Content outside the dataset might tend to revert towards the base Flux.1 Dev style.

• Avoid including conflicting artistic style prompts in your generation, as this can negatively impact the LoRA's ability to apply the Borderlands look.

---

Training Details

• Epochs: 100

• Dataset Size: Over 150+ Borderlands images

• Batch Size: 4

• Gradient Accumulation: 4 (Aids in generating varied outputs from similar prompts)

---

Usage Tips

• Experiment with the LoRA weight to find the intensity that best suits your image. Lower weights will be more subtle, higher weights more pronounced.

• Focus your prompts on subjects and scenes that align with the typical Borderlands universe for the best results.

• While this LoRA is trained for Flux.1 Dev, feel free to experiment cautiously with other Flux variants, but results may vary.

---

License/Terms of Use

This LoRA is provided under the FLUX.1 [dev] Non-Commercial License.

This means you are free to use, copy, modify, and distribute this LoRA, subject to the conditions outlined in the license. Please be mindful of the use restrictions, which generally prohibit using the model to intentionally create or disseminate illegal or harmful content.