--- license: apache-2.0 language: - en pipeline_tag: text-to-image tags: - stable-diffusion - alimama-creative library_name: diffusers --- # SD3 ControlNet Inpainting ![SD3](images/sd3.png)
a woman wearing a white jacket, black hat and black pants is standing in a field, the hat writes SD3
![bucket_alibaba](images/bucket_ali.png )
a person wearing a white shoe, carrying a white bucket with text "alibaba" on it
Finetuned controlnet inpainting model based on sd3-medium, the inpainting model offers several advantages: * Leveraging the SD3 16-channel VAE and high-resolution generation capability at 1024, the model effectively preserves the integrity of non-inpainting regions, including text. * It is capable of generating text through inpainting. * It demonstrates superior aesthetic performance in portrait generation. Compared with [SDXL-Inpainting](https://huggingface.co/diffusers/stable-diffusion-xl-1.0-inpainting-0.1) From left to right: Input image, Masked image, SDXL inpainting, Ours. ![0](images/0.png)
a tiger sitting on a park bench
![1](images/0r.png)
a dog sitting on a park bench
![2](images/1.png)
a young woman wearing a blue and pink floral dress
![3](images/3.png)
a woman wearing a white jacket, black hat and black pants is standing in a field, the hat writes SD3
![4](images/5.png)
an air conditioner hanging on the bedroom wall
# How to Use ``` python from diffusers.utils import load_image, check_min_version import torch # Local File from pipeline_sd3_controlnet_inpainting import StableDiffusion3ControlNetInpaintingPipeline, one_image_and_mask from controlnet_sd3 import SD3ControlNetModel check_min_version("0.29.2") # Build model controlnet = SD3ControlNetModel.from_pretrained( "alimama-creative/SD3-controlnet-inpainting", use_safetensors=True, ) pipe = StableDiffusion3ControlNetInpaintingPipeline.from_pretrained( "stabilityai/stable-diffusion-3-medium-diffusers", controlnet=controlnet, torch_dtype=torch.float16, ) pipe.text_encoder.to(torch.float16) pipe.controlnet.to(torch.float16) pipe.to("cuda") # Load image image = load_image( "https://huggingface.co/alimama-creative/SD3-Controlnet-Inpainting/blob/main/images/prod.png" ) mask = load_image( "https://huggingface.co/alimama-creative/SD3-Controlnet-Inpainting/blob/main/images/mask.jpeg" ) # Set args width = 1024 height = 1024 prompt="a woman wearing a white jacket, black hat and black pants is standing in a field, the hat writes SD3" generator = torch.Generator(device="cuda").manual_seed(48) input_dict = one_image_and_mask(image, mask, size=(width, height), latent_scale=pipe.vae_scale_factor, invert_mask = True) # Inference res_image = pipe( negative_prompt='deformed, distorted, disfigured, poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, mutated hands and fingers, disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, NSFW', prompt=prompt, height=height, width=width, control_image= input_dict['pil_masked_image'], # H, W, C, control_mask=input_dict["mask"] > 0.5, # B,1,H,W num_inference_steps=28, generator=generator, controlnet_conditioning_scale=0.95, guidance_scale=7, ).images[0] res_image.save(f'res.png') ``` ## Training Detail The model was trained on 12M laion2B and internal source images for 20k steps at resolution 1024x1024. * Mixed precision : FP16 * Learning rate : 1e-4 * Batch size : 192 * Timestep sampling mode : 'logit_normal' * Loss : Flow Matching ## Limitation Due to the fact that only 1024*1024 pixel resolution was used during the training phase, the inference performs best at this size, with other sizes yielding suboptimal results. We will initiate multi-resolution training in the future, and at that time, we will open-source the new weights.