search

autonomousvision / monosdf

[NeurIPS'22] MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstruction
MIT License
572 stars 53 forks source link

The program freezes when importing a model #88

Open 520jz opened 1 year ago

520jz commented 1 year ago

Hi, thanks for this fantastic work. When I trained the model using the scannet dataset, I had a strange issue where the program froze at 107 lines when running monosdf_train.py file image

After using debug to go in, you will find that the program has been stuck in the import (module) image

This is the output information of the terminal image

No matter how long I wait, the program will not run any further The conf file follows the default settings

The command line is as follows： CUDA_VISIBLE_DEVICES=0 python -m torch.distributed.launch --nproc_per_node 1 --nnodes=1 --node_rank=0 training/exp_runner.py --conf confs/scannet_mlp.conf --scan_id 9

The input data structure is as follows image

please, I need your help! thank you very much

niujinshuchong commented 1 year ago

Hi, we use our own implementation of hash-encoding so it will try to compile it for the first time when you run it which might takes some time. Here is the setup code: https://github.com/autonomousvision/monosdf/blob/main/code/hashencoder/backend.py

Can you also try to delete the tmp_build and rerun the training command again?