Example video

[autolife-nz]

Hello,

Thanks for making this code available - amazing work. We are currently trying to debug our installation but are having difficulty knowing if errors are from config, or from the input video file that we recorded ourselves. It would be very helpful to have one input video that we know works, so we can rule out input data as a cause of the errors. Would it be possible to release one of the example videos on the project page?

Many thanks.

[snow-nene]

Hello,

Thanks for making this code available - amazing work. We are currently trying to debug our installation but are having difficulty knowing if errors are from config, or from the input video file that we recorded ourselves. It would be very helpful to have one input video that we know works, so we can rule out input data as a cause of the errors. Would it be possible to release one of the example videos on the project page?

Many thanks.

Yes, I am encountering the same problem now.

[canyueduhong]

Yes, I am encountering the same problem now. The error is reported as follows

[ITER 30000] Saving Checkpoint [08/11 16:26:44]

Training complete. [08/11 16:26:45]
creating the SMPLX Decoder
Data path: /home/tom/fssd/GaussianHaircut/scene/folder
Fitted camera path: /home/tom/fssd/GaussianHaircut/scene/folder/3d_gaussian_splatting/stage1/cameras/30000_matrices.pkl
Overwriting COLMAP cameras with optimized cameras
Traceback (most recent call last):
File "/home/tom/fssd/GaussianHaircut/ext/NeuralHaircut/src/multiview_optimization/fit.py", line 83, in
main(
File "/home/tom/fssd/GaussianHaircut/ext/NeuralHaircut/src/multiview_optimization/fit.py", line 25, in main
dataset = Multiview_dataset(data_path, fitted_camera_path, use_scale, fixed_images, **conf['dataset'], device=conf['general.device'], batch_size=batch_size)
File "/home/tom/fssd/GaussianHaircut/ext/NeuralHaircut/src/multiview_optimization/dataset/dataset.py", line 162, in init
self.lmks = torch.from_numpy(np.stack([lmks[i] for i in self.good_views]))
File "<array_function internals>", line 200, in stack
File "/opt/conda/envs/gaussian_splatting_hair/lib/python3.9/site-packages/numpy/core/shape_base.py", line 460, in stack
raise ValueError('need at least one array to stack')
ValueError: need at least one array to stack

[autolife-nz]

Hello Yes we are stuck on the same issue 'ValueError('need at least one array to stack')', and not sure if it's a config/code issues, or data issue. Please email me at ***@***.*** and we can share any progress we make.

…

On Fri, 8 Nov 2024, 9:49 pm canyueduhong, ***@***.***> wrote: Yes, I am encountering the same problem now. The error is reported as follows [ITER 30000] Saving Checkpoint [08/11 16:26:44] Training complete. [08/11 16:26:45] creating the SMPLX Decoder Data path: /home/tom/fssd/GaussianHaircut/scene/folder Fitted camera path: /home/tom/fssd/GaussianHaircut/scene/folder/3d_gaussian_splatting/stage1/cameras/30000_matrices.pkl Overwriting COLMAP cameras with optimized cameras Traceback (most recent call last): File "/home/tom/fssd/GaussianHaircut/ext/NeuralHaircut/src/multiview_optimization/fit.py", line 83, in main( File "/home/tom/fssd/GaussianHaircut/ext/NeuralHaircut/src/multiview_optimization/fit.py", line 25, in main dataset = Multiview_dataset(data_path, fitted_camera_path, use_scale, fixed_images, **conf['dataset'], device=conf['general.device'], batch_size=batch_size) File "/home/tom/fssd/GaussianHaircut/ext/NeuralHaircut/src/multiview_optimization/dataset/dataset.py", line 162, in *init* self.lmks = torch.from_numpy(np.stack([lmks[i] for i in self.good_views])) File "<*array_function* internals>", line 200, in stack File "/opt/conda/envs/gaussian_splatting_hair/lib/python3.9/site-packages/numpy/core/shape_base.py", line 460, in stack raise ValueError('need at least one array to stack') ValueError: need at least one array to stack — Reply to this email directly, view it on GitHub <#13 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AANYE7ZEYMXE47XKPRVSHBTZ7R3JVAVCNFSM6AAAAABQ7IXVB2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRUGE2TAMJSGI> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[canyueduhong]

Okay, but your email is hidden

[autolife-nz]

Check our github profile (make sure you are logged in) and you should see my email address in profile. Seems Github removes emails in messages

…

On Fri, 8 Nov 2024, 11:13 pm canyueduhong, ***@***.***> wrote: Okay, but your email is hidden — Reply to this email directly, view it on GitHub <#13 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AANYE73KFAN6JZKLBH3UC73Z7SFDNAVCNFSM6AAAAABQ7IXVB2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRUGMYTENBWGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[canyueduhong]

ok i see

[canyueduhong]

Brother, I found out what the problem is. The problem is that when filtering high-quality pictures, setting it to 0.6, which causes no one to filter and the array is empty. Just change it to a lower level.

[autolife-nz]

Will try tomorrow. Thank you!

…

On Fri, 8 Nov 2024, 11:32 pm canyueduhong, ***@***.***> wrote: Brother, I found out what the problem is. The problem is that when filtering high-quality pictures, setting it to 0.6, which causes no one to filter and the array is empty. Just change it to a lower level. image.png (view on web) <https://github.com/user-attachments/assets/b0d8c351-e673-48e3-8d0e-2866d493e369> — Reply to this email directly, view it on GitHub <#13 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AANYE7Z4UWZIKRLM5XDT2ATZ7SHNHAVCNFSM6AAAAABQ7IXVB2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDINRUGM2TQMZTGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[autolife-nz]

Hello, we changed if score > 0.2 to 0.2, 0.1, 0.0 but we are still having the same problem. Did you get the whole process to work? Did you managed to get a final output? Also will send you a email. Thanks.

[canyueduhong]

I have successfully obtained the final result. We can continue our communication via email.