The QLoRA fine-tuning process of florence_pugh_lora_flux_nf4 takes inspiration from this post (https://huggingface.co/blog/flux-qlora). The training was executed on a local computer with 1000 timesteps and the same parameters as the link mentioned above, which took around 6 hours on 8GB VRAM 4060. The peak VRAM usage was around 7.7GB. To avoid running low on VRAM, both transformers and text_encoder were quantized. All the images generated here are using the below parameters

Height: 512
Width: 512
Guidance scale: 5
Num inference steps: 20
Max sequence length: 512
Seed: 0

Usage

python

import torch
from diffusers import FluxPipeline, FluxTransformer2DModel
from transformers import T5EncoderModel

text_encoder_4bit = T5EncoderModel.from_pretrained(
    "hf-internal-testing/flux.1-dev-nf4-pkg", subfolder="text_encoder_2",torch_dtype=torch.float16,)

transformer_4bit = FluxTransformer2DModel.from_pretrained(
        "hf-internal-testing/flux.1-dev-nf4-pkg", subfolder="transformer",torch_dtype=torch.float16,)

pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.float16,
                                    transformer=transformer_4bit,text_encoder_2=text_encoder_4bit)

pipe.load_lora_weights("je-suis-tm/florence_pugh_lora_flux_nf4",
                       weight_name='pytorch_lora_weights.safetensors')

prompt="close-up portrait of an edgy street style model named Florence Pugh with neon-colored eye makeup, focusing on intense black eyeshadow, glowing orange collar of a high-tech wear jacket visible, cyberpunk-inspired hairstyle with subtle colored highlights, background showing blurred city lights at night, piercing gaze directly at the camera, skin has a cool blue tint contrasting with warm lighting from below, small red digital elements floating near the face, photorealistic --ar 9:16 --quality 2 --style raw --v 6. 1"

image = pipe(
            prompt,
            height=512,
            width=512,
            guidance_scale=5,
            num_inference_steps=20,
            max_sequence_length=512,
            generator=torch.Generator("cpu").manual_seed(0),            
        ).images[0]

image.save("florence_pugh_lora_flux_nf4.png")

Trigger words

You should use Florence Pugh to trigger the image generation.

Download model

Download them in the Files & versions tab.