[evaluate.py] error : unrecognized data path

[HAMA-DL-dev]

Hello, I am a student who is interested in sensor fusion.

To run SCADC-DepthCompletion, I downloaded the data as it was posted on README.md and executed the following code.

However, there was an error that did not recognize the test data path even though the data file was downloaded normally.

So I tried to set up a number of data paths to solve this problem, but I couldn't.

I'd like some advice on how to solve this problem.

By the way, your repositories is very interesting for me, who wants to go to graduate school in the field of computer vision.

I also run CFCNet repository you uploaded, got the same image as the result in the repository and I read the paper to study the model structure.

The reason for running CFCNet and SCADC-DeepthCompletion is to learn more about deep learning models.

As a student who dreams of becoming a self-driving car's vision SW engineer, I am very impressed your repositories.

I hope you have a good day today, and thank you for reading the git issue!

[aimshirley]

Hello, I am a student who is interested in sensor fusion.
I am curious about how you deal with the H5.PY file. The DATASETS format of the H5.PY file I synthesized by myself is the same as yours, but when running evaluated.py, the result RMSE is NAN.I don't know what the problem is.Hope to get your reply,and thanks for your sharing!

[choyingw]

Hello, I am a student who is interested in sensor fusion.

To run SCADC-DepthCompletion, I downloaded the data as it was posted on README.md and executed the following code.

However, there was an error that did not recognize the test data path even though the data file was downloaded normally.

So I tried to set up a number of data paths to solve this problem, but I couldn't.

I'd like some advice on how to solve this problem.

By the way, your repositories is very interesting for me, who wants to go to graduate school in the field of computer vision.

I also run CFCNet repository you uploaded, got the same image as the result in the repository and I read the paper to study the model structure.

The reason for running CFCNet and SCADC-DeepthCompletion is to learn more about deep learning models.

As a student who dreams of becoming a self-driving car's vision SW engineer, I am very impressed your repositories.

I hope you have a good day today, and thank you for reading the git issue!

Sorry for the late reply. Give a space before ''--test_path" can fix the problem. (My typo in the readme)
Sample command is like python3 evaluate.py --name kitti --checkpoints_dir './test_ckpt' --test_path ./data

[choyingw]

Hello, I am a student who is interested in sensor fusion. I am curious about how you deal with the H5.PY file. The DATASETS format of the H5.PY file I synthesized by myself is the same as yours, but when running evaluated.py, the result RMSE is NAN.I don't know what the problem is.Hope to get your reply,and thanks for your sharing!

I downloaded given val set in the readme and run
python3 evaluate.py --name kitti --checkpoints_dir './test_ckpt' --test_path ./data
I didn't get NAN in the eval prompt

[aimshirley]

Hello, I am a student who is interested in sensor fusion. I am curious about how you deal with the H5.PY file. The DATASETS format of the H5.PY file I synthesized by myself is the same as yours, but when running evaluated.py, the result RMSE is NAN.I don't know what the problem is.Hope to get your reply,and thanks for your sharing!

I downloaded given val set in the readme and run python3 evaluate.py --name kitti --checkpoints_dir './test_ckpt' --test_path ./data I didn't get NAN in the eval prompt

It's my honor to get your reply. I want to convert my pictures into H5PY files to run this network. May I ask if there is any relevant code, because the H5PY conversion code I wrote by myself is very strange.