• Juveria Shafi Ahmed Shaikh
• Neha Jaiswar

Artificial Intelligence (AI) has transformed image generation through diffusion models like Stable Diffusion, DALL·E, and Midjourney. The FLUX series, created by Black Forest Labs, enhances AI-driven image synthesis with two models—FLUX.1 [dev] and FLUX.1 [schnell]—designed to optimize quality and speed. This study analyzes these models in terms of inference efficiency, output quality, and adaptability, focusing on LoRA (Low-Rank Adaptation) fine-tuning using the ostris/flux-dev-lora-trainer. Experimental findings reveal the trade-offs between FLUX.1 [dev]’s high-resolution output and FLUX.1 [schnell]’s faster processing. Additionally, the research explores LoRA’s role in improving model customization, providing insights into cost-efficient AI image generation techniques.

Flux, Image, Model, AI, Fine-Tuning, Quality, Speed, LoRA, Training, AI Image Generation.

"A STUDY ON FLUX-BASED MODELS AND LORA TRAINING TECHNIQUES: OPTIMIZING AI IMAGE GENERATION", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 2, page no.g446-g455, February-2025, Available :http://www.jetir.org/papers/JETIR2502651.pdf

"A STUDY ON FLUX-BASED MODELS AND LORA TRAINING TECHNIQUES: OPTIMIZING AI IMAGE GENERATION", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 2, page no. ppg446-g455, February-2025, Available at : http://www.jetir.org/papers/JETIR2502651.pdf