eth3d test error :shape mismatch

[xugzh96]

Hi, thanks for your great work!

I encountered a shape mismatch error when running priorda test with eth3d datasets:

RuntimeError: shape mismatch: value tensor of shape [77271060] cannot be broadcast to indexing result of shape [12423]

Here is my command:
export HF_ENDPOINT=https://hf-mirror.com && /home/yj/anaconda/envs/priorda/bin/priorda test --image_path /home/yj/eth3d/allDataset/pipes/images/00000000.jpg --prior_path /home/yj/eth3d/allDataset/pipes/mono_depths/00000000.jpg --pattern 1000 --visualize 1

My image size is 4141 × 6220. The error occurs at sparse_sampler.py line 221:
sparse_depth[down_mask] = prior.flatten()

Could you please advise:

1. Is this caused by my input format ?
2. Or is it because my image size (4141×6220) is too large and causes dimension mismatch?

Thank you!