ERROR: failed to create sparse model in `ns-process-data`

[fanzz1208]

Describe the bug Running COLMAP bundle adjustment... (This may take a while)────────────────────────────────────────────── 💀 💀 💀 ERROR 💀 💀 💀 ─────────────────────────────────────────────── Error running command: colmap mapper --database_path /workspace/nerfdata/068/colmap/database.db --image_path /workspace/nerfdata/068/images --output_path /workspace/nerfdata/068/colmap/sparse --Mapper.ba_global_function_tolerance 1e-6 ──────────────────────────────────────────────────────────────────────────────────────────────────────────────────────── ERROR: failed to create sparse model

[audiocommander]

I stumbled upon this error, too. I'm on a MacBook Air M2 AARCH64, PyTorch with mps support for silicon. I tried installing COLMAP by downloading a precompiled release, which runs, but I couldn't make it work for python scripts. So I installed COLMAP via brew and it seems to kinda work. Now I got exactly this error, too.

[tudorjnu]

I have the same error. Any solutions found so far?

[camolina2]

I have this error too :( any solutions?

[rexainn]

same problem

[AlexanderVoldby]

I am also getting this error, I'm just not getting the "ERROR: failed to create sparse model" message. Have you tried running this with the --verbose flag to see extended logs? I personally don't get any extra information from this though...

[rexainn]

I thought it was because the dataset i used didn't have enough views...

[AlexanderVoldby]

Did you fix the error by using more video data?

[rexainn]

I used different videos for different views, so I got this error. Once I changed one video like shooting arround the same scene, the error is fixed.

[AlexanderVoldby]

Okay, I am currently just using a single video but maybe I was moving the camera too fast, I guess I’ll try with another video

[cici19850]

Okay, I am currently just using a single video but maybe I was moving the camera too fast, I guess I’ll try with another video

I am also getting this error, I'm just not getting the "ERROR: failed to create sparse model" message too.Did you fix the error?

[AryamanSharma17]

Not sure if this is helpful but for me, CUDA was out of memory. I was able to get feature matching and sparse point cloud by closing other memory-consuming scripts.

[Tiangj9178]

Have you ever encountered this situation? How to deal with it? nerfstudio) root@autodl-container-8b8a4382f2-a9434246:/# ns-process-data images --data /root/autodl-tmp/mtree3-305/images --output-dir /nerfstudio/data1/mtree3 [16:13:59] 🎉 Done copying images with prefix 'frame_'. process_data_utils.py:340 [16:15:38] 🎉 Done extracting COLMAP features. colmap_utils.py:137 [16:23:40] 🎉 Done matching COLMAP features. colmap_utils.py:151 [16:39:49] 🎉 Done COLMAP bundle adjustment. colmap_utils.py:173 🎉 Done refining intrinsics. colmap_utils.py:184 einsum(): the number of subscripts in the equation (2) does not match the number of process_data.py:526 dimensions (1) for operand 1 and no ellipsis was given