search

yindaz / DeepCompletionRelease

Deep Depth Completion of a Single RGB-D Image
http://deepcompletion.cs.princeton.edu/
563 stars 141 forks source link

Looking for a copy of data #45

Open yindaz opened 3 years ago

yindaz commented 3 years ago

Bad news to announce: Our technic staff accidentally deleted all the shared data related to this paper and repo. We are still in the process of trying to recover the data. Meanwhile, please let me (yindaz@gmail.com) know if you happen to downloaded the data in whole or part. Thanks!

kuersatp commented 3 years ago

Hi Yinda, there's a paper that proposes to pretrain their network on the data that you published: https://sites.google.com/view/cleargrasp and https://arxiv.org/pdf/1910.02550.pdf . I think that they might still have the data. Maybe you can get into contact with them. It would be great as we also rely on the data that you published. Best, Nick

stevenhxy commented 2 years ago

Hi @yindaz

Sorry for bothering. May I ask you that the dataset is recovered or not? The website always shows "Forbidden, You don't have permission to access this resource." when I download from https://github.com/yindaz/DeepCompletionRelease/tree/master/data and https://github.com/yindaz/DeepCompletionRelease/tree/master/pre_train_model. Is there any method to download the dataset and pre-train model?

Thanks Steven

Darker-zq commented 2 years ago

I'm excited to learn your approach. And can I get your datasets now? Thanks. Zhang

hebangwen commented 1 year ago

Sorry to bother you after so long. Have you found the realsense.zip dataset? The link to realsense.zip is still forbidden.

petekk-93 commented 1 year ago

I contacted the authors of the "ClearGrasp..." paper. They do not have the data either but wanted to send me the pre-trained weights at least. However, I did not obtain them either. One proposal from their side was to use Apple's recent synthetic indoor dataset for pre-training. Maybe that would work.