search

zzangjinsun / NLSPN_ECCV20

Park et al., Non-Local Spatial Propagation Network for Depth Completion, ECCV, 2020
MIT License
321 stars 55 forks source link

A question about the evaluation on pre-trained model on KITTI DC datasets? #43

Closed Danica-zc closed 2 years ago

Danica-zc commented 2 years ago

hello，thanks for sharing the code! I downloaded the pre-trained model on KITTI DC datasets from here (https://drive.google.com/file/d/11by_1oglcncSHFeF3S2ldcysQYjyDcQh/view). After testing, it tells "KeyError:""filename""storage" not found". I think the pre-trained model may be corrupted .

zzangjinsun commented 2 years ago

Hi, from my side it works fine. Please download it again or be sure that you are using python3.

Danica-zc commented 2 years ago

yes, I am sure I am using python3.7

Danica-zc commented 2 years ago

There are someting wrong still. Is it right? NLSPN_KITTI_DC.pt 102637kb

zzangjinsun commented 2 years ago

Exactly. I even tested with the newest (stable) version of pytorch and apex on anaconda and it worked fine.

Danica-zc commented 2 years ago

Did you use the model downloaded from this link ?(https://drive.google.com/file/d/11by_1oglcncSHFeF3S2ldcysQYjyDcQh/view)

zzangjinsun commented 2 years ago

Yes, I downloaded it from the released link and tested it on a new computer but it worked fine. If it is available, please try the file on a different computer.

Danica-zc commented 2 years ago

python main.py --dir_data PATH_TO_KITTI_DC --data_name KITTIDC --split_json ../data_json/kitti_dc.json \ --patch_height 240 --patch_width 1216 --gpus 0,1,2,3 --max_depth 90.0 --num_sample 0 \ --test_only --pretrain ../results/NLSPN_KITTI_DC.pt --preserve_input --save NAME_TO_SAVE --legacy is it the input of this command, no need to change anything else？