Tsung-Han (Patrick) Wu

I met the same problem too. After a lot of trial-and-error, I found that the following scripts are able to run successfully. ```bash python3 train.py configs/img/onet.yaml python3 train.py configs/pointcloud/onet.yaml python3...

Hi, - For the pertained model, please see [here](https://github.com/tsunghan-mama/Depth-Completion/blob/master/doc/train_test.md). - For the training and testing data, please see [here](https://github.com/tsunghan-mama/Depth-Completion/blob/master/doc/data.md). Thanks.

Please see the latest issue (#23) .

Hi, we have released our dataset. Please follow in instructions described on our README to obtain dataset access. Thanks. Best, Tsung-Han Wu

Hi, the training and testing pair data comes from "[Deep Depth Completion of a Single RGB-D Image](https://openaccess.thecvf.com/content_cvpr_2018/papers/Zhang_Deep_Depth_Completion_CVPR_2018_paper.pdf)".

Hi, Please follow the [instructions](https://github.com/tsunghan-mama/Depth-Completion/blob/master/doc/train_test.md) to modify the dataset path and then you could read the file successfully. Best, Tsung-Han Wu

Hi, Yes as you've mentioned, we indeed use the number of observed pixels in evaluation. Our implementation can be shown [here](https://github.com/tsunghan-mama/Depth-Completion/blob/master/depth_completion/eval.py#L53). We first calculate the number of unobserved pixels and...

Hi, Thanks for your great findings! It seems that our implementation might be incorrect. However, I don't think it would affect the comparison results between these depth completion methods. Sincerely...

Hi everyone, we have released our dataset. Please follow the instructions described on our README to obtain dataset access. Thanks. Best, Tsung-Han Wu

Hi @eecoder-dyf, Sorry for the inconvenience and I have updated the permission. Thanks for your kind reminder. Best, Tsung-Han