IndexError: list index out of range

[chensh1127]

(octree-gs) I run the code: bash single_train.sh

the output is: Setting up [LPIPS] perceptual loss: trunk [vgg], v[0.1], spatial [off] /home//miniconda3/envs/octree-gs/lib/python3.8/site-packages/torchvision/models/_utils.py:208: UserWarning: The parameter 'pretrained' is deprecated since 0.13 and will be removed in 0.15, please use 'weights' instead. warnings.warn( /home//miniconda3/envs/octree-gs/lib/python3.8/site-packages/torchvision/models/_utils.py:223: UserWarning: Arguments other than a weight enum or None for 'weights' are deprecated since 0.13 and will be removed in 0.15. The current behavior is equivalent to passing weights=VGG16_Weights.IMAGENET1K_V1. You can also use weights=VGG16_Weights.DEFAULT to get the most up-to-date weights. warnings.warn(msg) Loading model from: /home//miniconda3/envs/octree-gs/lib/python3.8/site-packages/lpips/weights/v0.1/vgg.pth not found tf board 2024-04-05 13:12:23,960 - INFO: args: Namespace(add_color_dist=False, add_cov_dist=False, add_level=False, add_opacity_dist=False, appearance_dim=0, appearance_lr_delay_mult=0.01, appearance_lr_final=0.0005, appearance_lr_init=0.05, appearance_lr_max_steps=40000, base_layer=-1, checkpoint_iterations=[], coarse_factor=1.5, coarse_iter=5000, compute_cov3D_python=False, data_device='cuda', debug=False, debug_from=-1, densify_grad_threshold=0.0002, detect_anomaly=False, dist2level='round', dist_ratio=0.999, ds=1, eval=True, extend=1.1, extra_ratio=0.5, extra_up=0.01, feat_dim=32, feature_lr=0.0075, fork=2, gpu='-1', images='images', init_level=-1, ip='127.0.0.1', iterations=40000, lambda_dssim=0.2, levels=-1, min_opacity=0.005, mlp_color_lr_delay_mult=0.01, mlp_color_lr_final=5e-05, mlp_color_lr_init=0.008, mlp_color_lr_max_steps=40000, mlp_cov_lr_delay_mult=0.01, mlp_cov_lr_final=0.004, mlp_cov_lr_init=0.004, mlp_cov_lr_max_steps=40000, mlp_featurebank_lr_delay_mult=0.01, mlp_featurebank_lr_final=1e-05, mlp_featurebank_lr_init=0.01, mlp_featurebank_lr_max_steps=40000, mlp_opacity_lr_delay_mult=0.01, mlp_opacity_lr_final=2e-05, mlp_opacity_lr_init=0.002, mlp_opacity_lr_max_steps=40000, model_path='outputs/data/truck/baseline/2024-04-05_13:12:21', n_offsets=10, offset_lr_delay_mult=0.01, offset_lr_final=0.0001, offset_lr_init=0.01, offset_lr_max_steps=40000, opacity_lr=0.02, percent_dense=0.01, port=22315, position_lr_delay_mult=0.01, position_lr_final=0.0, position_lr_init=0.0, position_lr_max_steps=40000, progressive=True, quiet=False, random_background=False, ratio=1, resolution=-1, resolution_scales=[1.0], rotation_lr=0.002, save_iterations=[-1], scaling_lr=0.007, source_path='data/data/truck', start_checkpoint=None, start_stat=500, success_threshold=0.8, test_iterations=[-1], undistorted=False, update_anchor=True, update_from=1500, update_interval=100, update_ratio=0.2, update_until=20000, use_feat_bank=False, use_wandb=False, visible_threshold=0.9, warmup=False, white_background=False) [10000, 20000, 30000, 40000] [10000, 20000, 30000, 40000] Backup Finished! 2024-04-05 13:12:24,129 - INFO: Optimizing outputs/data/truck/baseline/2024-04-05_13:12:21 Output folder: outputs/data/truck/baseline/2024-04-05_13:12:21 [05/04 13:12:24] Tensorboard not available: not logging progress [05/04 13:12:24] Traceback (most recent call last): File "train.py", line 560, in training(lp.extract(args), op.extract(args), pp.extract(args), dataset, args.test_iterations, args.save_iterations, args.checkpoint_iterations, args.start_checkpoint, args.debug_from, wandb, logger) File "train.py", line 89, in training scene = Scene(dataset, gaussians, ply_path=ply_path, shuffle=False, logger=logger, resolution_scales=dataset.resolution_scales) File "/home/Octree-GS/scene/init.py", line 51, in init scene_info = sceneLoadTypeCallbacks["City"](args.source_path, args.random_background, args.white_background, args.eval, args.ds, undistorted=args.undistorted) File "/home/Octree-GS/scene/dataset_readers.py", line 381, in readCityInfo json_path = glob.glob(os.path.join(path, f".json"))[0].split('/')[-1] IndexError: list index out of range**

I also tried my custom data and the SfM data sets for Tanks&Temples and Deep Blending are hosted by 3D-Gaussian-Splatting. the error is the same.

[tongji-rkr]

Usually this happens because the path to the data is wrong, that is, the source_path is not set correctly. I encourage you to check train.sh, I see data/data/truck in your error, which should be the cause of the problem.

[chensh1127]

you are right. I have modified my dataset path, but now I am getting an error: nvrtc: error: invalid value for --gpu-architecture (-arch） my coda version is 11.6 python version is 3.8 torch is 1.12.1 Has anyone encountered this before?

[tongji-rkr]

I guess it is caused by the cuda driver, and what GPU do you use to train?

[chensh1127]

[tongji-rkr]

You can try lowering your python version, I'm using 3.7.13.

[levelzhang]

I also encountered the same problem（nvrtc: error: invalid value for --gpu-architecture (-arch））, how did you solve it

[tongji-rkr]

Please let ue know where the error is in the.py file. We currently find some similar problems in 4090.

[tongji-rkr]

There are two possible solutions:

1. conda install pytorch==2.0.1 torchvision==0.15.2 torchaudio==2.0.2 pytorch-cuda=11.8 -c pytorch -c nvidia
2. 

[cat-fish-yu]

you are right. I have modified my dataset path, but now I am getting an error: nvrtc: error: invalid value for --gpu-architecture (-arch）你是对的.我已经修改了我的数据集路径，但现在我得到一个错误：nvrtc：error：invalid value for --gpu-architecture（-arch） my coda version is 11.6 我的版本是11.6 python version is 3.8 Python版本3.8 torch is 1.12.1  火炬是1.12.1 Has anyone encountered this before? 以前有人遇到过吗？

Hi, have you solved this problem?

[tongji-rkr]

You can refer to the two solutions I have given, and we have identified the cause of the problem