search

XuyangBai / D3Feat.pytorch

[PyTorch] Official Implementation of CVPR'20 oral paper - D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features https://arxiv.org/abs/2003.03164
138 stars 20 forks source link

train and test the network #3

Closed david-ngo-sixense closed 3 years ago

david-ngo-sixense commented 3 years ago

Thanks for your work. It is very interesting. Can you explain how to test the network? I run: python test.py and the error: FileNotFoundError: [Errno 2] No such file or directory: '/data/D3Feat/snapshot//config.json'

And how did you train the network ? It sounds that the file train.py is missing. Thanks.

XuyangBai commented 3 years ago

training_3DMatch.py is for training

For testing: python test.py --chosen_snapshot [timestr of the model]

david-ngo-sixense commented 3 years ago

Thanks for your quick reply. I have this error when training: Traceback (most recent call last): File "training_3DMatch.py", line 128, in dist_type=config.dist_type, AttributeError: 'EasyDict' object has no attribute 'dist_type'

You did not define the argument for 'dis_type' in config.py.

I do not understand the testing. For example, when I run: python test.py --chosen_snapshot 15 I have the error: FileNotFoundError: [Errno 2] No such file or directory: '/data/D3Feat/snapshot/15/config.json' The file config.json is created after training ? I download the dataset 3DMatch following your tensorflow respository. And there is no config.json file.

XuyangBai commented 3 years ago

After you train the model, there will be a folder in snapshot and the folder name is what you used for testing. The config.json is also generated automatically in this folder

python test.py --chosen_snapshot [the foldername]