Can you please let me know can I generate the image that is provided in (b)?
Can you please share the code that is used for visualization/rendering the output results like that

Hi,

I'm trying to download the pre-trained model with the provided script:

wget https://cmu.box.com/s/6bq16zba09rjfezfpflfhepgse0ja9cf

but it didn't return anything, I'm wondering if this link is broken?

Thanks

Hi, Nilesh, Thanks for your work especially for the experiment part. I'm trying to reproduce your work. However, I got stuck in the NYU data part. Specifically, I found the object_obj folder of your project is not provided.

After some code checking, I figured out this object mat file names are the same as the work
Support Surface Prediction in Indoor Scenes e.g. bed_14.mat.
However the content of this object is not the same format of your code.
['surfaces', 'comp', '__version__', '__header__', '__globals__']

Hello,

It seems like the pretrained model files (*.pth) provided in your cachedir.tar.gz download do not match the names specified in your training examples. I am specifically looking for the box3d_base_spatial_mask_common_upsample and dwr_base_spatial_mask_common_upsample specified here for training on NYU.

Thanks

Hi,

Thank you for the great work.

I tried to deploy the environment and run the codes, but most of the necessary dependent resources seem not downloadable, like all files under 'https://people.eecs.berkeley.edu/~shubhtuls/'.

Besides, there are some small issues like: pip cannot locate PyMesh2 of version 0.2.1 and yyatg (yatg?)

Could you please provide some help with it? Many thanks.

https://github.com/nileshkulkarni/relative3d/blob/master/docs/installation.md#setup-virtualenv

conda env create -f docs/condenv.yml

should be

conda env create -f docs/condaenv.yml

Hello, Nil, Could you share all the models for the results on the table? Specifically, for your reproducted baseline results. I tried to use batch size 24 for baseline reproduction, but can not reproduct your factored3d result. From your link here (now is broken), I can only obtain the following model.
box3d_base_spatial_mask_common_upsample/ dwr_base_spatial_mask_common_upsample/ layout_pred/ nyu_relnet_dwr_pos_ft/ object_autoenc_32/ pretrained_dwr_shape_ft/ suncg_relnet_dwr_pos_ft/
However, from your baseline guideline here, your reproducted baseline models are
box3d_base_factored3d_baseline, dwr_base_factored3d_baseline_ft, box3d_base_factored3d_baseline_nyu, dwr_base_factored3d_baseline_nyu
which are missing.

Hello,

This line requires .png image files from the NYUv2 dataset but the website does not provide png files. Where do you get these .png files or how do you generate them?