Use multiple controlnets tensor size error

### Describe the bug

I get a wrong tensor size error when trying to run multiple controlnets

### Reproduction

```
import argparse
import os
import random

import cv2
import numpy as np
import torch
from PIL import Image
from diffusers import (
AutoencoderKL,
EulerAncestralDiscreteScheduler, )
from diffusers.utils import load_image
from torchvision import transforms

from replace_bg.model.controlnet import ControlNetModel
from replace_bg.model.pipeline_controlnet_sd_xl import StableDiffusionXLControlNetPipeline
from replace_bg.utilities import resize_image, remove_bg_from_image, paste_fg_over_image, get_control_image_tensor

import torch

from transformers import DPTFeatureExtractor, DPTForDepthEstimation, DPTImageProcessor

depth_estimator = DPTForDepthEstimation.from_pretrained("Intel/dpt-hybrid-midas").to("cuda")
feature_extractor = DPTImageProcessor.from_pretrained("Intel/dpt-hybrid-midas")
def get_depth_map(image):
image = feature_extractor(images=image, return_tensors="pt").pixel_values.to("cuda")
with torch.no_grad(), torch.autocast("cuda"):
depth_map = depth_estimator(image).predicted_depth
depth_map = torch.nn.functional.interpolate(
depth_map.unsqueeze(1),
size=(1024, 1024),
mode="bicubic",
align_corners=False,
)
depth_min = torch.amin(depth_map, dim=[1, 2, 3], keepdim=True)
depth_max = torch.amax(depth_map, dim=[1, 2, 3], keepdim=True)
depth_map = (depth_map - depth_min) / (depth_max - depth_min)
image = torch.cat([depth_map] * 3, dim=1)
image = image.permute(0, 2, 3, 1).cpu().numpy()[0]
image = Image.fromarray((image * 255.0).clip(0, 255).astype(np.uint8)).resize((1024,1024))
return image

def run_sdxl_multicontrolnet(args, prompt, negative_prompt):
cd = ControlNetModel.from_pretrained(
"diffusers/controlnet-depth-sdxl-1.0",
torch_dtype=torch.float16, variant="fp16")
cc = ControlNetModel.from_pretrained("diffusers/controlnet-canny-sdxl-1.0", torch_dtype=torch.float16,variant="fp16")
controlnets = [cd,cc]
vae = AutoencoderKL.from_pretrained("madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16)
pipe = StableDiffusionXLControlNetPipeline.from_pretrained(
'stabilityai/stable-diffusion-xl-base-1.0',
controlnet=controlnets, torch_dtype=torch.float16, variant="fp16",vae=vae)

pipe.to("cuda")
pipe.enable_model_cpu_offload()

image = load_image(args.image_path)
image = resize_image(image)
image.save('./tmp.jpg')
depth_image = get_depth_map(image)

low_threshold, high_threshold = 100, 200
input_image = cv2.Canny(image, low_threshold, high_threshold)
input_image = input_image[:, :, None]
input_image = np.concatenate([input_image, input_image, input_image], axis=2).astype(np.uint8)
canny_image = Image.fromarray(input_image)

assert depth_image.size==canny_image.size
images = [depth_image.resize((1024, 1024)),canny_image.resize((1024, 1024))]
gen_img = pipe(
negative_prompt=negative_prompt,
prompt=prompt ,
controlnet_conditioning_scale=[1.0,1.0],
num_inference_steps=30,
image=images,

).images[0]
gen_img.save('gen.png')
depth_image.save('depth.png')
canny_image.save('canny.png')

```

### Logs

```shell
0%| | 0/30 [00:01
run_sdxl_multicontrolnet(args, args.prompt, args.negative_prompt)
File "/home/ilias.papastratis/workdir/bria_models/run_bria_modelsv2.py", line 227, in run_sdxl_multicontrolnet
gen_img = pipe(
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
return func(*args, **kwargs)
File "/home/ilias.papastratis/workdir/bria_models/replace_bg/model/pipeline_controlnet_sd_xl.py", line 1377, in __call__
down_block_res_samples, mid_block_res_sample = self.controlnet(
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/accelerate/hooks.py", line 166, in new_forward
output = module._old_forward(*args, **kwargs)
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/diffusers/pipelines/controlnet/multicontrolnet.py", line 48, in forward
down_samples, mid_sample = controlnet(
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1511, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/home/ilias.papastratis/workdir/envs/object_env/lib/python3.10/site-packages/torch/nn/modules/module.py", line 1520, in _call_impl
return forward_call(*args, **kwargs)
File "/home/ilias.papastratis/workdir/bria_models/replace_bg/model/controlnet.py", line 801, in forward
sample = sample + controlnet_cond
RuntimeError: The size of tensor a (128) must match the size of tensor b (1024) at non-singleton dimension
```

### System Info

diffusers==0.26.0
Python=3.10

### Who can help?

_No response_