Sign In

Y2K Asian Aphrodite

34

Updated: Jul 31, 2026

character

Download

1 variant available

SafeTensor

158.03 MB

Verified:

Type
LoRA
Stats

40

0

Reviews
Published

Jul 31, 2026

Base Model

Flux.2 Klein 9B-base

Training
Steps: 5,000
Epochs: 20
Usage Tips
Strength: 1
Hash
AutoV2
9AB92F7CAB
Trigger Words
Y2K Asian Aphrodite
default creator card background decoration
Reactions - 9591

9.6K

Followers - 1186

1.2K

Downloads - 3738

3.7K

Created on Civitai

The FLUX.1 [dev] Model is licensed by Black Forest Labs. Inc. under the FLUX.1 [dev] Non-Commercial License. Copyright Black Forest Labs. Inc.

IN NO EVENT SHALL BLACK FOREST LABS, INC. BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH USE OF THIS MODEL.

💫 Y2K Asian Aphrodite – Early 2000s Aesthetic LoRA

Trained on a curated dataset of 200 early 2000s aesthetic photographs, this LoRA captures the golden era of early 2000s film photography — natural East Asian women, cool-toned cinematic lighting, authentic film grain and texture, layered Y2K jewelry, and era-accurate fashion from that decade. Perfect for outdoor lifestyle portraits, editorial fashion shots, cinematic full-body compositions, and candid film-style imagery with exceptional photorealism.

Sister model to Y2K Brunette Aphrodite and Y2K Blonde Aphrodite — same aesthetic DNA, same training pipeline, optimized for East Asian facial structure and dark hair types.

⚙️ Recommended Settings:

  • CFG Scale: 3.5 – 5.0

  • Steps: 20 – 30

  • Sampler: Euler or DPM++ 2M

  • Strength: 0.75 – 0.85

  • Trigger Word: y2kasianaphrodite

Prompt Tips:

  • Add film grain, 35mm photography, kodak portra, chromatic aberration for authentic Y2K film feel

  • Add early 2000s fashion, low rise jeans, crop top, layered necklaces, cargo pants for era-accurate styling

  • Add soft natural lighting, golden hour, shallow depth of field, outdoor portrait for best compositional results

  • Keep facial structure prompts specific — East Asian features, dark straight hair, sharp cheekbones prevents the model from drifting toward other ethnicities in the dataset spread

  • Avoid heavy makeup tags — some model versions tend to invent strong makeup unprompted

⚠️ Note: Results and strengths vary by base model version. Portrait detail is strongest on FLUX Klein-based versions. Cinematic film grain and moody atmosphere are strongest on the Z-Image Turbo version. Check individual version notes for model-specific tips and recommended settings.

Enjoy the model, and please leave a like or share your favorite generations!