StableDiffusionXLAdapterPipeline 多个adapter无法工作

[jttqbc666]

这是我的代码： `from diffusers import StableDiffusionXLAdapterPipeline, T2IAdapter, EulerAncestralDiscreteScheduler, AutoencoderKL,MultiAdapter from diffusers.utils import load_image, make_image_grid from controlnet_aux.midas import MidasDetector import torch from controlnet_aux.canny import CannyDetector

device='cuda:0'

load adapter

depth_midas_adapter = T2IAdapter.from_pretrained( "TencentARC/t2i-adapter-depth-midas-sdxl-1.0", torch_dtype=torch.float16, varient="fp16" ).to(device) canny_adapter = T2IAdapter.from_pretrained("TencentARC/t2i-adapter-canny-sdxl-1.0", torch_dtype=torch.float16, varient="fp16").to(device) adapters = MultiAdapter([depth_midas_adapter,canny_adapter])

load euler_a scheduler

model_id = 'stabilityai/stable-diffusion-xl-base-1.0' euler_a = EulerAncestralDiscreteScheduler.from_pretrained(model_id, subfolder="scheduler") vae=AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16) pipe = StableDiffusionXLAdapterPipeline.from_pretrained( model_id, vae=vae, adapter=adapters, scheduler=euler_a, torch_dtype=torch.float16, variant="fp16", ).to(device) pipe.enable_xformers_memory_efficient_attention()

midas_depth = MidasDetector.from_pretrained( "valhalla/t2iadapter-aux-models", filename="dpt_large_384.pt", model_type="dpt_large" ).to("cuda") canny_detector = CannyDetector() url = "https://huggingface.co/Adapter/t2iadapter/resolve/main/figs_SDXLV1.0/org_mid.jpg" image = load_image(url) depth_midas_control_image = midas_depth( image, detect_resolution=512, image_resolution=1024 ) canny_control_image = canny_detector(image, detect_resolution=384, image_resolution=1024)#.resize((1024, 1024)) prompt = "A photo of a room, 4k photo, highly detailed" negative_prompt = "anime, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured" control_weights=[0.6,0.6]

gen_images = pipe( prompt=prompt, negative_prompt=negative_prompt, image=[[depth_midas_control_image,canny_control_image]], num_inference_steps=30, adapter_conditioning_scale=control_weights, guidance_scale=7.5, ).images[0] gen_images.save('out_mid.png')`

报错：AttributeError: 'MultiAdapter' object has no attribute 'total_downscale_factor'