neuralsim
neuralsim copied to clipboard
neuralsim: 3D surface reconstruction and simulation based on 3D neural rendering.
I am trying the streetsurf using waymo datasets. It seems that they all use lotd for encoding, in the code, I cannot find the link between the lotd and NGP....
Thanks for the great work! I encountered the following problem when I tried to run the script withmask_withlidar_withnormal.240219.yaml: ``` Traceback (most recent call last): File "code_single/tools/train.py", line 43, in from...
Hi, thanks for the code release of multi objects, and I have downloaded it and want to have a try. But I encountered a problem, which is as follows. Could...
首先感谢作者无私奉献开源了这个repo,我尝试了几个方法,目前Neuralsim是效果最好的。 我基于waymo做了18组实验(这些场景是在repo提供的81个动态场景里挑选的,有低速行驶和高速行驶的),有7组的loss不收敛。 使用的config: [all_occ.with_normals.240201.yaml](https://github.com/PJLab-ADG/neuralsim/blob/main/code_multi/configs/exps/fg_neus%3Dpermuto/all_occ.with_normals.240201.yaml)。 使用的segmentation模型:https://github.com/open-mmlab/mmsegmentation/tree/main/configs/mask2former  想请教的问题: 1. loss不收敛问题 1. 自车速度高时loss不收敛是什么原因?对于自车高速行驶的场景有哪些超参数需要调整?config默认的配置segment-9653249092自车是低速行驶的,从实验结果来看这组效果最好,但是我们的业务更关注高速的场景。 2. 重建效果不好问题有哪些优化思路?(视频见下文) 1. 车道线不清晰 2. 空中多了一块云/水滴,可能是什么问题导致的,有哪些优化的思路? 3. 彻底糊了的原因,见seg-364414视频 4. 最开始几帧远处闪了几下 具体复现结果如下: # loss不收敛导致训到一半挂了(7组),自车速度在70km/h左右 ``` # - segment-1758724094753801109_1251_037_1271_037_with_camera_labels # 64km/h...
基于lidar和nomask的config,目前跑出来一个模型,看了下形状大致对的,能看出大概路面形状,车道线和灯杆形状,但是mesh渲染的颜色完全不对,感觉渲染完全错了。 我使用的数据集,有两个相机(前后视),目前只用了前视相机,斜装激光装在车后面扫描,所以为了对齐前视相机,我会把前几秒的激光数据对齐当前相机这一帧一起处理; 目前有啥经验可以指点吗,比如调整什么参数,或者调整lidar数据,还是有什么输入数据给的有问题?
I am trying to train using a custom dataset on v0.6.0. I have setup my dataset according to autonomous_driving.md but when I try training I get the following NotImplementedError: ...
Great work! When will the code of objects and scenes editing in streetsurf be released ?Very much looking forward to it!
Firstly, great work and thanks for making it open-source. I setup everything following the readme for both streetsurf and nr3d. I wanted to use the withmask_nolidar.240219.yaml config file, made the...
Hi, I'm trying to use monodepth supervision to train the multi-object model, but there are some problems like this, which seems to be related to ImagePatchDataset. Could you give me...
Thanks for the great work. The demos I have seen from the project website are all quite short (within 10 seconds) and travel only limited distance. Have you tried to...