zju3dv/easyvolcap

[SIGGRAPH Asia 2023 (Technical Communications)] EasyVolcap: Accelerating Neural Volumetric Video Research

easyvolcap's Issues

RuntimeError: cudaErrorUnknown(999): unknown error

I run according to the readme documentation
this step：Running 3DGS+T
runevc -t gui -c configs/exps/gaussiant/gaussiant_${expname}.yaml,configs/specs/superm.yaml
encounter RuntimeError: cudaErrorUnknown(999): unknown error

Please help me.

easyvolcap.models.samplers.point_planes_sampler very slow

it takes nearly 2hours on v100 to finish this part, is this different on architecture 75+ gpus？

Bad result in Neural3DV dataset

I use Neural3DV dataset sear_steak, follow the dataset conversion scripts in neural3dv_to_easyvolcap.py to generate yml files, and train I3mhet use config in configs/exps/l3mhet/l3mhet_sear_steak.yaml, the result is so bad which val psnr is only 6.36

2023-12-19 10:44:27.029948 easyvolcap.runners.evaluators.volumetric_video_evaluator -> summarize:                                             volumetric_video_evaluator.py:79
                           {                                                                                                                                                  
                               'psnr_mean': 6.362307601504856,                                                                                                                
                               'psnr_std': 0.16340760990474038,                                                                                                               
                               'ssim_mean': 0.018442549,                                                                                                                      
                               'ssim_std': 0.0059690047,                                                                                                                      
                               'lpips_mean': 0.676931189166175,                                                                                                               
                               'lpips_std': 0.010196555702566121                                                                                                              
                           }   

error happened when inference enerfi

after training 3DGS-T, I try to Inference ENeRFi , and the surfs dirs is generated by 3DGS-T at the begin of training

an error occurred

*** 'ImageBasedInferenceDataset' object has no attribute 'mks_bytes'

it seems enerfi need get mask bytes of every frame?