strkyyy/anime-90s-retro

90s Retro Anime VHS SDXL LoRA

What it does

Capture the essence of 90s retro anime aesthetic with this LoRA, featuring a soft pastel palette, grainy VHS texture, slight chromatic aberration, hand-painted cel look with visible brushwork on backgrounds, and a warm nostalgic mood.

How to use

Trigger the style using the word '4n1m390s' in your prompt. Suggested strength is 0.85 for optimal results; feel free to adjust for personal preference. Works best with EMA sampler at 20 steps.

Best for

• Scenes set during sunset or twilight hours
• Rainy day settings
• Schoolyard scenes of everyday life
• Magical girl transformations and spell casting
• Futuristic robot battles in neon-lit arenas

Tips

• Combine with other LoRAs for a unique twist on the retro anime theme.
• Adjust strength to blend the 90s aesthetic seamlessly into your scene.

Example Prompts

4n1m390s -- Portrait of a young boy with tousled hair, wearing a red hoodie and jeans, standing on an old wooden bridge, warm sunset lighting

4n1m390s -- Nighttime scene of a bustling city street lined with retro arcade games and neon signs, soft pastel colors blending into the night sky

4n1m390s -- Close-up of a vintage Walkman player on a wooden desk, warm ambient lighting casting shadows, slightly grainy texture adds to nostalgic feel

4n1m390s -- A young girl in a flowing dress running through a field with a kite flying high above her head, soft pastel colors and slight chromatic aberration

4n1m390s -- Warm nostalgic mood of an old bookshelf filled with VHS tapes, hand-painted cel look backgrounds with visible brushwork, warm lighting

4n1m390s -- Abstract pattern inspired by 90s anime aesthetic, soft pastel colors and grainy texture, slight chromatic aberration adds to the retro feel

Trigger word

Activate with 4n1m390s. Recommended strength: 0.9.

Showcase

Usage with diffusers

from diffusers import AutoPipelineForText2Image
import torch

pipe = AutoPipelineForText2Image.from_pretrained("stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16).to("cuda")
pipe.load_lora_weights("strkyyy/anime-90s-retro")
image = pipe("4n1m390s, your prompt", num_inference_steps=30).images[0]