Getting “Sizes of tensors must match” error when using ComfyUI WanVideoWrapper (wan2.2) to generate video

Question

I am trying to generate a video using Wan 2.2. My goal is to take a motion sequence from an input video and a single reference image, and then generate a new video where the character in the reference image performs the actions from the input video.For this, I am using this [workflow][1] provided in ComfyUI-WanVideoWrapper.

However, when I run the workflow, I consistently get the following error:

```bash

Sizes of tensors must match except in dimension 2.

Expected size 887 but got size 880 for tensor number 1 in the list.

```

I replaced the following node from the workflow:
```json
{
  "id": 73,
  "type": "DownloadAndLoadDepthAnythingV2Model",
  "pos": [-1430.5018, -396.4539],
  "size": [441, 82],
  "order": 24,
  "mode": 2,
  "outputs": [
    {
      "name": "da_v2_model",
      "type": "DAMODEL",
      "links": [82]
    }
  ],
  "properties": {
    "cnr_id": "comfyui-depthanythingv2",
    "ver": "003d7b44bafd3a8a4c3693a9ca3ddcd72f4883ab"
  }
}

with a DWPose node:

"156": {
  "inputs": {
    "detect_hand": "enable",
    "detect_body": "enable",
    "detect_face": "enable",
    "resolution": 512,
    "bbox_detector": "yolox_l.onnx",
    "pose_estimator": "dw-ll_ucoco_384_bs5.torchscript.pt",
    "scale_stick_for_xinsr_cn": "disable",
    "image": ["96", 0]
  },
  "class_type": "DWPreprocessor",
  "_meta": {
    "title": "DWPose Estimator"
  }
}

I already resized both the input video and the reference image to the recommended resolution of 480x832. The fps is set to 60 with a total frame count of 81, and the model files are placed according to the workflow’s requirements. Currently, the resolution in the DWpose node is set to 512.

Could this error be caused by the resolution setting in DWpose, or is there another preprocessing step in the workflow that I am missing?

傅靖茹 · Accepted Answer · 2025-09-30 05:32:42Z

1

I’ve encountered a similar issue before.
The reason is that the DWPose node in your workflow alters the image size, even if you’ve already resized the video and reference image to 480×832. The solution is to add a Resize Image node after the DWPose node to resize the output back to 480×832.
Hope this helps!

answered Sep 30 at 5:32

傅靖茹

511 silver badge2 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Getting “Sizes of tensors must match” error when using ComfyUI WanVideoWrapper (wan2.2) to generate video

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related