je-suis-tm/monica_bellucci_lora_flux

Monica Bellucci Lora Flux

Prompt

Training With QLoRA: Monica Bellucci with long, dark hair cascading in loose waves down her back. Her hair is parted on the side, framing her face. Dark almond-shaped eyes in heavy black eyeliner, adorned with thick eyelashes and thick arched brows, are accentuated by heavy smokey eye makeup, including false lashes and dark eyeshadow. A hint of a smile reveals full pink lips. She dons a black jacket over a pink top, suggesting a casual yet stylish ensemble. The image's natural lighting hints at an outdoor setting, with daytime vibes. Her poised stance indicates she is posing for the photo shoot.

Prompt

Training Without QLoRA: Monica Bellucci with long, dark hair cascading in loose waves down her back. Her hair is parted on the side, framing her face. Dark almond-shaped eyes in heavy black eyeliner, adorned with thick eyelashes and thick arched brows, are accentuated by heavy smokey eye makeup, including false lashes and dark eyeshadow. A hint of a smile reveals full pink lips. She dons a black jacket over a pink top, suggesting a casual yet stylish ensemble. The image's natural lighting hints at an outdoor setting, with daytime vibes. Her poised stance indicates she is posing for the photo shoot.

Prompt

Testing With QLoRA: portrait of a beautiful mermaid called Monica Bellucci. there are a lot of small goldfish around and against the background of marine life. the light barely breaks through the water column. Cinematic, 35mm, F/2. 8, Vignette, 4k, Spotlight, Cinematic Lighting, insanely detailed and intricate, elegant, ornate, super detailed, --ar 1:2 --q 2

Prompt

Testing Without QLoRA: portrait of a beautiful mermaid called Monica Bellucci. there are a lot of small goldfish around and against the background of marine life. the light barely breaks through the water column. Cinematic, 35mm, F/2. 8, Vignette, 4k, Spotlight, Cinematic Lighting, insanely detailed and intricate, elegant, ornate, super detailed, --ar 1:2 --q 2

All files are also archived in https://github.com/je-suis-tm/huggingface-archive in case this gets censored.

This a non-quantized version of https://huggingface.co/je-suis-tm/monica_bellucci_lora_flux_nf4. Both are trained on the same dataset. The training is based on https://github.com/huggingface/diffusers/blob/main/examples/dreambooth/test_dreambooth_lora_flux.py. The training took 3 hours on A100 80GB with max VRAM consumption at 35GB. The inference consumes 36GB of VRAM.

Train

    export MODEL_NAME="black-forest-labs/FLUX.1-dev"
    export INSTANCE_DIR="/pvol/monica_bellucci"
    export OUTPUT_DIR="/pvol/monica_bellucci_lora_flux"
    accelerate config default
    accelerate launch train_dreambooth_lora_flux1.py \
      --pretrained_model_name_or_path=$MODEL_NAME  \
      --mixed_precision="bf16" \
      --dataset_name=$INSTANCE_DIR \
      --output_dir=$OUTPUT_DIR \
      --gradient_checkpointing \
      --instance_prompt="Monica Bellucci" \
      --caption_column="text" \
      --resolution=1024 \
      --train_batch_size=1 \
      --guidance_scale=1 \
      --use_8bit_adam \
      --checkpointing_steps=100 \
      --gradient_accumulation_steps=4 \
      --optimizer="adamW" \
      --learning_rate=1e-4 \
      --lr_scheduler="constant" \
      --lr_warmup_steps=100 \
      --max_train_steps=1500 \
      --rank=4 \
      --seed="0" 

Usage

import torch
from diffusers import FluxPipeline

device = "cuda:0"

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

pipe.load_lora_weights("je-suis-tm/monica_bellucci_lora_flux",
                       weight_name='pytorch_lora_weights.safetensors')


prompt = "portrait of a beautiful mermaid called Monica Bellucci. there are a lot of small goldfish around and against the background of marine life. the light barely breaks through the water column. Cinematic, 35mm, F/2. 8, Vignette, 4k, Spotlight, Cinematic Lighting, insanely detailed and intricate, elegant, ornate, super detailed, --ar 1:2 --q 2"

image = pipe(
  prompt=prompt,
  generator=torch.Generator(device=device).manual_seed(42),
  num_inference_steps=50, # 28 is a good trade-off
  guidance_scale=4,
  height=1024, 
  width=1024,
).images[0]

image.save("monica_bellucci.png")

Trigger words

You should use Monica Bellucci to trigger the image generation.

Download model

Download them in the Files & versions tab.