graspnet / suctionnet-baseline Goto Github PK

Hello! Could you please point out the location of score_maps that are used during training? Where are they stored? I searched the repository and the website, but haven't found anything.

Also, that would be great if you could explain what they represent and how are they collected. Thanks in advance!

Provided dataset not working while training the model, please provide the latest training files for the same.

Hi, do you provide same annotation tool like graspnet for suctionnet?

In my implementation of the neural network described in your paper, the network predicts a center score map as a part of its output. However, when using my own dataset and modifying the depth map range to match the range in the dataset, I noticed that the predicted center score map is not performing optimally.

I would appreciate your guidance on how to optimize the center score map prediction of the neural network. Specifically, what are some techniques or approaches that could improve the accuracy of the predicted center score map?

Thank you for your time and expertise.

The center score map:

the mix map:

sampled:

smoothness:

ess:

in the file inference.py there are lots of flags mentioned at the beginning, what the meaning of these? and where to download the checkpoint?

Could you provide more detailed docs or tools for annotating suction poses?
Thanks!

In this work, which camera (ex. D435) was used to build the dataset called 'RealSense'?

Dear Author,
Can you send me relevant python files about evaluating the performance of trained network so that I can get the results just like Table4 in the paper? I have only seen inference.py in the repo. Thanks in advance and my email is [email protected].