what is wrong with inpaint background?


```py
import torch
import numpy as np
from PIL import Image, ImageOps
from diffusers import AutoPipelineForInpainting
import rembg

modelName = 'kandinsky-community/kandinsky-2-2-decoder-inpaint'

pipe = AutoPipelineForInpainting.from_pretrained(modelName, torch_dtype = torch.float16).to('cuda')
# pipe.enable_model_cpu_offload()

prompt = 'young lady, in front of garden, beautiful, 4k'

image = Image.open('./a.jpg')
mask = Image.open('./mask123.png').convert('RGB')
# mask = ImageOps.invert(mask)
# mask = mask.convert('L')
# mask = np.array(mask)

image = pipe(
image = image,
mask_image = mask,
prompt = prompt,
width = image.width,
height = image.height,
num_inference_steps = 50,
# prior_steps = 10,
).images[0]

display(image)
```