streetwear-core-lora

Streetwear Core LoRA

Version: 2.0 (Retrained June 2025)

A LoRA (Low-Rank Adaptation) fine-tuned model for generating streetwear fashion photography in the distinctive "streetwear core" aesthetic.

Latest Update: This model has been retrained with improved training techniques and optimized parameters for better quality and consistency.

Model Description

This LoRA adapter specializes in generating images with a streetwear aesthetic, featuring:

• Fashion Focus: Contemporary streetwear clothing including oversized tees, baggy jeans, sneakers, and accessories
• Brand Elements: Daily Paper clothing, Nike Air Force 1s, and other streetwear staples
• Photography Style: Editorial photography with film camera aesthetics, natural lighting, and urban settings
• Diverse Models: Both male and female models in various poses and settings
• 90s/2000s Influence: Captures the nostalgic feel of 90s hip-hop and early 2000s fashion

Trigger Word

Use streetwear_core as the trigger word to activate the LoRA's style.

Usage

import torch
from diffusers import FluxPipeline

# Load the base model
pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16)

# Load the LoRA
pipe.load_lora_weights("your-username/streetwear-core-lora")

# Generate image
prompt = "streetwear_core style, confident female model wearing olive green fitted tee and denim jeans, gold accessories, urban setting, natural daylight, editorial photography"

image = pipe(
    prompt,
    num_inference_steps=28,
    guidance_scale=3.5,
    width=1024,
    height=1024,
).images[0]

image.save("streetwear_output.png")

Example Prompts

• streetwear_core style, male model in teal matching t-shirt and shorts set, white sneakers, urban residential street backdrop, film camera aesthetic
• streetwear_core style, female model wearing camo print sweatsuit, confident pose, natural lighting, editorial style
• streetwear_core style, relaxed pose, oversized graphic tee and patterned baggy jeans, outdoor setting, 90s hip-hop fashion

Training Details

• Training Images: 29 high-quality streetwear fashion photographs
• Training Steps: 8 (as indicated by model filename)
• Base Model: FLUX.1-dev
• Resolution: Various resolutions optimized for editorial photography
• Concept: streetwear_core aesthetic

Training Dataset Characteristics

The model was trained on diverse streetwear imagery featuring:

• Male and female models of various ethnicities
• Urban and outdoor settings
• Natural lighting and film photography aesthetics
• Contemporary streetwear brands (Daily Paper, Nike, Adidas)
• Various poses: seated, standing, leaning, walking
• Accessories: jewelry, headwear, bags
• Color palettes: earth tones, pastels, urban colors

Recommended Settings

• Steps: 20-40
• CFG Scale: 3.5-7.0
• Resolution: 1024x1024 or 768x1024
• Sampler: DPM++ 2M Karras or Euler a

License

This model is licensed under the CreativeML OpenRAIL-M license. You are free to use this model for commercial and non-commercial purposes, but please review the full license terms.

Disclaimers

• This model may not work well for generating non-streetwear fashion styles
• Results may vary based on the prompt and generation parameters
• The model reflects the specific aesthetic of the training dataset

Credits

Training data curated and captioned with focus on streetwear fashion photography aesthetics.