par Shakker-Labs
Open source · 8k downloads · 490 likes
AWPortrait FL est un modèle d'IA spécialisé dans la génération d'images de portraits, affiné à partir de FLUX.1-dev avec un ensemble de données de haute qualité. Il excelle dans la composition, les détails et le réalisme, notamment pour la peau et les textures, offrant des résultats plus fins et plus naturels que son modèle de base. Conçu pour la photographie de mode, il convient particulièrement aux créateurs cherchant à produire des portraits esthétiques et professionnels. Ce qui le distingue, c'est sa capacité à générer des images d'une grande précision visuelle, tout en restant accessible via des versions optimisées comme le LoRA pour réduire l'utilisation mémoire.
AWPortrait-FL is finetuned on FLUX.1-dev using the training set of AWPortrait-XL and nearly 2,000 fashion photography photos with extremely high aesthetic quality. It has remarkable improvements in composition and details, with more delicate and realistic skin and textual. Trained by DynamicWang at AWPlanet.
The following example shows a simple comparison with FLUX.1-dev under the same parameter setting.
import torch
from diffusers import FluxPipeline
pipe = FluxPipeline.from_pretrained("Shakker-Labs/AWPortrait-FL", torch_dtype=torch.bfloat16)
pipe.to("cuda")
prompt = "close up portrait, Amidst the interplay of light and shadows in a photography studio,a soft spotlight traces the contours of a face,highlighting a figure clad in a sleek black turtleneck. The garment,hugging the skin with subtle luxury,complements the Caucasian model's understated makeup,embodying minimalist elegance. Behind,a pale gray backdrop extends,its fine texture shimmering subtly in the dim light,artfully balancing the composition and focusing attention on the subject. In a palette of black,gray,and skin tones,simplicity intertwines with profundity,as every detail whispers untold stories."
image = pipe(prompt,
num_inference_steps=24,
guidance_scale=3.5,
width=768, height=1024,
).images[0]
image.save(f"example.png")
To save memory, we also add a LoRA version to achieve same performance.
import torch
from diffusers import FluxPipeline
pipe = FluxPipeline.from_pretrained("black-forest-labs/FLUX.1-dev", torch_dtype=torch.bfloat16)
pipe.load_lora_weights('Shakker-Labs/AWPortrait-FL', weight_name='AWPortrait-FL-lora.safetensors')
pipe.fuse_lora(lora_scale=0.9)
pipe.to("cuda")
prompt = "close up portrait, Amidst the interplay of light and shadows in a photography studio,a soft spotlight traces the contours of a face,highlighting a figure clad in a sleek black turtleneck. The garment,hugging the skin with subtle luxury,complements the Caucasian model's understated makeup,embodying minimalist elegance. Behind,a pale gray backdrop extends,its fine texture shimmering subtly in the dim light,artfully balancing the composition and focusing attention on the subject. In a palette of black,gray,and skin tones,simplicity intertwines with profundity,as every detail whispers untold stories."
image = pipe(prompt,
num_inference_steps=24,
guidance_scale=3.5,
width=768, height=1024,
).images[0]
image.save(f"example.png")
You can also download this model at Shakker AI, where we provide an online interface to generate images.
This model is trained by our copyrighted users DynamicWang. We release this model under permissions. The model follows flux-1-dev-non-commercial-license and the generated images are also non commercial.