search

kcheng1021 / GaussianPro

[ICML2024] Official code for GaussianPro: 3D Gaussian Splatting with Progressive Propagation
https://kcheng1021.github.io/gaussianpro.github.io/
MIT License
551 stars 36 forks source link

FileNotFoundError: [Errno 2] No such file or directory: './cache/propagated_depth/costs.dmb' #38

Closed cokeshao closed 2 months ago

cokeshao commented 2 months ago

Thanks for you amazing work! But I have problems when running waymo.sh. The waymo.sh crashed due to FileNotFoundError. Could you please tell me what it is and how to solve This problem. Thanks a lot.

(3dgs) xxx@M603:~/3dgs/GaussianPro$ bash ./scripts/waymo.sh                                                                                   
Optimizing $                                                                                                                                       
Output folder: $ [26/04 15:40:05]                                                                                                                  
Tensorboard not available: not logging progress [26/04 15:40:05]                                                                                   
Reading camera 198/198 [26/04 15:40:06]                                                                                                            
Loading Training Cameras [26/04 15:40:06]                                                                                                          
[ INFO ] Encountered quite large input images (>1.6K pixels width), rescaling to 1.6K.                                                             
 If this is not desired, please explicitly specify '--resolution/-r' as 1 [26/04 15:40:06]                                                         
Loading Test Cameras [26/04 15:40:30]                                                                                                              
Number of points at initialisation :  72378 [26/04 15:40:34]                                                                                       
Training progress:   0%|                                                                                              | 0/30000 [00:00<?, ?it/s]   
[ITER 1] Evaluating test: L1 0.2074004739522934 PSNR 11.672923622131348 [26/04 15:40:34]                                                           

[ITER 1] Evaluating train: L1 0.23484502732753754 PSNR 10.758105468750001 [26/04 15:40:34]                                                         

[ITER 1] Saving Gaussians [26/04 15:40:34]                                                                                                         
Training progress:   7%|████▍                                                              | 2000/30000 [01:42<21:08, 22.07it/s, Loss=0.0258575]   
[ITER 2000] Evaluating test: L1 0.022805737778544426 PSNR 29.186945571899415 [26/04 15:42:17]                                                      

[ITER 2000] Evaluating train: L1 0.018611638620495797 PSNR 30.844982147216797 [26/04 15:42:17]                                                     
Training progress:  23%|███████████████▋                                                   | 7000/30000 [05:26<17:53, 21.42it/s, Loss=0.0220800]   
[ITER 7000] Evaluating test: L1 0.014742725938558578 PSNR 32.64846481323242 [26/04 15:46:00]                                                       

[ITER 7000] Evaluating train: L1 0.011671697162091732 PSNR 34.663641357421874 [26/04 15:46:00]                                                     

[ITER 7000] Saving Gaussians [26/04 15:46:00]                                                                                                      
Training progress:  60%|███████████████████████████████████████▊                          | 18070/30000 [14:24<09:28, 20.98it/s, Loss=0.0208992]Tra
ining progress:  60%|███████████████████████████████████████▊                          | 18100/30000 [14:25<09:24, 21.07it/s, Loss=0.0245535]Traini
ng progress: 100%|██████████████████████████████████████████████████████████████████| 30000/30000 [23:39<00:00, 21.14it/s, Loss=0.0163919]         

[ITER 30000] Evaluating test: L1 0.011243130750954152 PSNR 35.02509567260742 [26/04 16:04:13]                                                      

[ITER 30000] Evaluating train: L1 0.008955066464841366 PSNR 37.520050048828125 [26/04 16:04:13]                                                    

[ITER 30000] Saving Gaussians [26/04 16:04:13]                           

Training complete. [26/04 16:04:17]
Looking for config file in $/cfg_args                                    
Config file found: $/cfg_args
Rendering $                                                                                                                                        
Loading trained model at iteration 30000 [26/04 16:04:24]                                                                                          
Reading camera 198/198 [26/04 16:04:25]                                                                                                            
Loading Training Cameras [26/04 16:04:25]                                                                                                          
[ INFO ] Encountered quite large input images (>1.6K pixels width), rescaling to 1.6K.                                                             
 If this is not desired, please explicitly specify '--resolution/-r' as 1 [26/04 16:04:25]                                                         
Loading Test Cameras [26/04 16:04:49]                                                                                                              
Rendering progress: 100%|█████████████████████████████████████████████████████████████████████████████████████| 173/173 [05:48<00:00,  2.02s/it]   
Rendering progress: 100%|███████████████████████████████████████████████████████████████████████████████████████| 25/25 [00:50<00:00,  2.03s/it]   

Scene: $                                                                                                                                           
Method: ours_30000                                                                                                                                 
Metric evaluation progress:   0%|                                                                                        | 0/25 [00:00<?, ?it/s]Dow
nloading: "https://download.pytorch.org/models/vgg16-397923af.pth" to xxx/.cache/torch/hub/checkpoints/vgg16-397923af.pth               
100%|████████████████████████████████████████████████████████████████████████████████████████████████████████| 528M/528M [01:01<00:00, 8.96MB/s]   
Downloading: "https://raw.githubusercontent.com/richzhang/PerceptualSimilarity/master/lpips/weights/v0.1/vgg.pth" to xxx/.cache/torch/hu
b/checkpoints/vgg.pth               

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████
| 7.12k/7.12k [00:01<00:00, 4.14kB/s]                                    
Metric evaluation progress: 100%|███████████████████████████████████████████████████████████████████████████████| 25/25 [04:46<00:00, 11.46s/it]███
| 7.12k/7.12k [00:01<00:00, 4.14kB/s]                                    
  SSIM :    0.9500440                                                    
  PSNR :   34.9467659                                                    
  LPIPS:    0.2342081                                                    

Optimizing $                                                             
Output folder: $ [26/04 16:16:33]                                        
Tensorboard not available: not logging progress [26/04 16:16:33]                                                                                   
Reading camera 198/198 [26/04 16:16:34]                                  
Loading Training Cameras [26/04 16:16:35]                                
[ INFO ] Encountered quite large input images (>1.6K pixels width), rescaling to 1.6K.                                                             
 If this is not desired, please explicitly specify '--resolution/-r' as 1 [26/04 16:16:35]                                                         
Loading Test Cameras [26/04 16:16:58]                                    
Number of points at initialisation :  72378 [26/04 16:17:02]                                                                                       
Training progress:   0%|            
           | 0/30000 [00:00<?, ?it/s]                                    
[ITER 1] Evaluating test: L1 0.2074004739522934 PSNR 11.672923622131348 [26/04 16:17:02]                                                           

[ITER 1] Evaluating train: L1 0.23484502732753754 PSNR 10.758105468750001 [26/04 16:17:02] 

[ITER 1] Saving Gaussians [26/04 16:17:02]
Training progress:   3%|███▌                                                                                                       | 1010/30000 [00:50<22:57, 21.04it/s, Loss=0.3880446]sh: 1: ./submodules/Propagation/Propagation: not found
Traceback (most recent call last):
  File "xxx/3dgs/GaussianPro/train.py", line 372, in <module>
    training(lp.extract(args), op.extract(args), pp.extract(args), args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from)
  File "xxx/3dgs/GaussianPro/train.py", line 129, in training
    propagated_depth, cost, normal = read_propagted_depth('./cache/propagated_depth')
  File "xxx/3dgs/GaussianPro/utils/general_utils.py", line 244, in read_propagted_depth
    cost = readDepthDmb(os.path.join(path, 'costs.dmb'))
  File "xxx/3dgs/GaussianPro/utils/general_utils.py", line 192, in readDepthDmb
    inimage = open(file_path, "rb")
FileNotFoundError: [Errno 2] No such file or directory: './cache/propagated_depth/costs.dmb'
Training progress:   3%|███▌                                                                                                       | 1010/30000 [00:50<24:23, 19.81it/s, Loss=0.3880446]
Looking for config file in $/cfg_args
Config file found: $/cfg_args
Rendering $
Loading trained model at iteration 30000 [26/04 16:17:58]
Reading camera 198/198 [26/04 16:17:59]
Loading Training Cameras [26/04 16:17:59]
[ INFO ] Encountered quite large input images (>1.6K pixels width), rescaling to 1.6K.
 If this is not desired, please explicitly specify '--resolution/-r' as 1 [26/04 16:17:59]
Loading Test Cameras [26/04 16:18:23]
Rendering progress: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 173/173 [05:18<00:00,  1.84s/it]
Rendering progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 25/25 [00:46<00:00,  1.85s/it]

Scene: $
Method: ours_30000
Metric evaluation progress: 100%|███████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 25/25 [00:27<00:00,  1.09s/it]
  SSIM :    0.9500440
  PSNR :   34.9467659
  LPIPS:    0.2342081

my waymo.sh

python train.py -s /home/xxx/3dgs/dataset/Waymo/segment-102751 -m $save_path$ \
                --eval --position_lr_init 0.000016 --scaling_lr 0.001 --percent_dense 0.0005 --port 1021 --dataset waymo 

python render.py -m $save_path$
python metrics.py -m $save_path$

python train.py -s /home/xxx/3dgs/dataset/Waymo/segment-102751 -m $save_path$ \
                --eval --flatten_loss --position_lr_init 0.000016 --scaling_lr 0.001 --percent_dense 0.0005 --port 1021 --dataset waymo \
                --sky_seg --normal_loss --depth_loss --propagation_interval 30 --depth_error_min_threshold 0.8 --depth_error_max_threshold 1.0 \
                --propagated_iteration_begin 1000 --propagated_iteration_after 12000 --patch_size 20 --lambda_l1_normal 0.001 --lambda_cos_normal 0.001

python render.py -m $save_path$
python metrics.py -m $save_path$
cokeshao commented 2 months ago

I thought that-m $save_path$would create a folder with a random name under /output to save data like the 3dgs does. But it creates a file called '$' in the current directory and saves it. It's my fault.