mengmengliu1998 / gatraj Goto Github PK

Thank you for your excellent work. Hello, I have some problems with visualization. If possible, can you share the source code?

Thank for your outstanding work!

1. global_embed = self.multihead_proj_global(global_embed).view(-1, self.num_modes, self.hidden_size) # [N, F, D]

What is the purpose of this function? Why design this function？

2. Here

the parameters of this function are time tensor x_encoded_dense and spatial tensor hidden_state_global

but in

the time tensor's name change to global_embed and the spatial tensor’s name change to local_embed, Is this correct？

Thank you!

Thank you for your excellent work. Can you please provide the nuScenes dataset?

Thank you for your outstanding work !

There is an option called --input_offset in the program.

When this option is set to True, the train_x is the difference between batch_norm_gt , batch_norm_gt is the difference between absolute position. why do we do this?

Hi, @mengmengliu1998 I have had the pleasure of reading your paper, I feel it is very well written, now I want to study the code of the paper, I see you have provided the code on ETH/UCY dataset, I would like to ask if you are convenient to provide the code on nuScenes dataset, if it is convenient can you send it to my email, [email protected],

Hello, this code cannot find the relevant Nuscenes data preprocessing file. May I ask how to execute it on the Nuscenes dataset? Thank you.

Hello, I have recently read your article and your proposed framework is excellent. I would like to further understand how the model is applied to the nuScene dataset? I would like to ask if you are convenient to provide the code on nuScenes dataset. Thank you！ This is my email: [email protected]

hello，I have a question regarding the implementation of mode k and the likelihood π in your algorithm.

I would like to know whether mode k is applied to individual agents independently, or is it used to describe all agents within the entire scene? In othjer words, are the likelihoods π associated with modes k the same for all agents, or does each agent have distinct likelihoods π for their respective modes k?

looking forward to your reply！！ thank you！