proj_matrix was not set

JOP-Lee / READ

AAAI2023，implementation of "READ: Large-Scale Neural Scene Rendering for Autonomous Driving", the experimental results are significantly better than Nerf-based methods

https://github.com/JOP-Lee/READ-Large-Scale-Neural-Scene-Rendering-for-Autonomous-Driving

GNU General Public License v2.0

447 stars 55 forks source link

proj_matrix was not set #31

Closed vinodrajendran001 closed 1 year ago

vinodrajendran001 commented 1 year ago

Hi,

I'm trying to reproduce the results for kitti dataset with the headless mode but I am getting the following error

May I know how to resolve it?

Thanks.

JOP-Lee commented 1 year ago

@Sylvia6

wuzuyin commented 1 year ago

me too

wuzuyin commented 1 year ago

@Sylvia6

what do you mean?

Sylvia6 commented 1 year ago

If in the headess mode, during training, you should not use OpenGL in dynamic.py to realize point rasterization. Instead, the code will set input_ to value {} (which will be calculated by out implemented renderer in train.py）. In this way, we can avoid rasterization perfomed every time in the __getitem__ , thus gain acceleration. Here's is the specific code details:

In READ/src/READ/datasets/dynamic.py:

While in READ/src/train.py:

Sylvia6 commented 1 year ago

@wuzuyin @vinodrajendran001 Would you please take a look at headless mode, is the code related to headless args works?