je-suis-tm/brie_larson_lora_flux_nf4

Brie Larson Lora Flux NF4

Prompt

Training With QLoRA: Brie Larson in a striking superhero costume is depicted in the image. She has shoulder-length, tousled hair, neatly parted on one side to reveal her ear, with a few strands softly framing her face. Her costume is a vibrant combination of red and blue, featuring a high collar and a prominent star emblem on the chest. Gold accents highlight the edges of the suit, adding a touch of regal flair. The pose and overall presentation convey a sense of readiness and anticipation, suggesting she's poised for action, whether it be taking flight or confronting a challenge. The image evokes a feeling of power and preparedness, hinting at a heroic mission about to unfold.

Prompt

Training Without QLoRA: Brie Larson in a striking superhero costume is depicted in the image. She has shoulder-length, tousled hair, neatly parted on one side to reveal her ear, with a few strands softly framing her face. Her costume is a vibrant combination of red and blue, featuring a high collar and a prominent star emblem on the chest. Gold accents highlight the edges of the suit, adding a touch of regal flair. The pose and overall presentation convey a sense of readiness and anticipation, suggesting she's poised for action, whether it be taking flight or confronting a challenge. The image evokes a feeling of power and preparedness, hinting at a heroic mission about to unfold.

Prompt

Testing With QLoRA: young egyptian queen named Brie Larson in super exposing pharoah clothes and necklace, background in ancient egyptian palace, hyper realistic, skin imperfections, flirty eyes, ultra details

Prompt

Testing Without QLoRA: young egyptian queen named Brie Larson in super exposing pharoah clothes and necklace, background in ancient egyptian palace, hyper realistic, skin imperfections, flirty eyes, ultra details

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

The QLoRA fine-tuning process of brie_larson_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

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/brie_larson_lora_flux_nf4",
                       weight_name='pytorch_lora_weights.safetensors')

prompt="young egyptian queen named Brie Larson in super exposing pharoah clothes and necklace, background in ancient egyptian palace, hyper realistic, skin imperfections, flirty eyes, ultra details"

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("brie_larson_lora_flux_nf4.png")

Trigger words

You should use Brie Larson to trigger the image generation.

Download model

Download them in the Files & versions tab.