FLUX-REALISM

FLUX-REALISM is an experimental, advanced image generation application designed to provide highly realistic image synthesis workflows. Powered by the state-of-the-art FLUX architecture, the suite offers a dual-model workspace featuring FLUX.1-Krea-dev and a specialized FLUX.1-dev pipeline integrated with a realism-enhancing LoRA (strangerzonehf/Flux-Super-Realism-LoRA). The application provides an interactive web interface optimized with custom quality presets, negative prompt expansion rules, and on-the-fly VAE decoding switches (incorporating madebyollin/taef1 for preview stages and a high-fidelity Krea VAE for final selection frames). Fully GPU-accelerated and capable of yielding images up to 2048x2048, FLUX-REALISM acts as a powerful testing suite for researchers and creators exploring photorealistic text-to-image boundaries.

Key Features

• Dual-Model Inference: Seamlessly alternate between the fast, iteration-friendly flux.1-krea-dev pipeline and the hyper-detailed flux.1-dev-realism pipeline equipped with a dedicated realism LoRA adapter.
• Custom Quality Style Presets: Built-in resolution and prompt styling configurations (e.g., Ultra HD 4K, 8K Cinematic) that automatically rewrite user prompts to maximize environmental detail, photorealism, and lifelike clarity.
• Iterative Image Streaming: Features a specialized flux_pipe_call_that_returns_an_iterable_of_images implementation that hooks into the pipeline's denoising loop to deliver continuous generation feedback.
• Granular Generation Controls: Complete control over generation dimensions (up to 2048px), inference steps, seed randomization, and guidance scale settings.
• Automated Batch Export: Supports packaging multiple generated images into a single downloadable ZIP archive instantly on the frontend.

Repository Structure

├── app.py
├── LICENSE
├── pre-requirements.txt
├── pyproject.toml
├── README.md
├── requirements.txt
└── uv.lock

Installation and Requirements

To run FLUX-REALISM locally, configure a Python environment with the following dependencies. A compatible CUDA-enabled GPU with sufficient VRAM is required to execute FLUX.1 models.

Standard PIP Installation

1. Update pip:

pip install pip>=26.1.1

2. Install dependencies:

pip install -r requirements.txt

Running with uv (Recommended)

uv is an extremely fast Python package and project manager written in Rust, which ensures rapid environment setup and complete reproducibility.

Step 1 — Install uv

• macOS / Linux: curl -LsSf https://astral.sh/uv/install.sh | sh
• Windows: powershell -c "irm https://astral.sh/uv/install.ps1 | iex"

Step 2 — Clone the repository

git clone https://github.com/PRITHIVSAKTHIUR/FLUX-REALISM.git
cd FLUX-REALISM

Step 3 — Initialize the project and install dependencies

uv sync

Step 4 — Run the script

uv run app.py

---

Core Requirements List

The application depends on the following primary packages (defined in requirements.txt):

git+https://github.com/huggingface/accelerate.git
git+https://github.com/huggingface/diffusers.git
git+https://github.com/huggingface/peft.git
sentencepiece
transformers
safetensors
torch==2.11.0
torchvision
requests
kernels
hf_xet
spaces
pillow
gradio
av

Usage

Once the application is running, open your browser to the local address provided in your terminal (typically http://127.0.0.1:7860/).

1. Select Model: Choose your base model pipeline from the dropdown menu (flux.1-krea-dev or flux.1-dev-realism).
2. Enter Prompt: Describe your scene inside the main prompt text field (e.g., "Ultra realistic cinematic portrait of a woman standing in neon rain").
3. Configure Options: Expand the "Additional Options" accordion to modify inference parameters, choose quality styles, toggle negative prompts, or enable ZIP compilation.
4. Generate: Click "Generate Images" to execute. The output gallery will display results, along with specific execution time and final seed metadata.

License and Source

• License: Apache-2.0 license
• GitHub Repository: https://github.com/PRITHIVSAKTHIUR/FLUX-REALISM