Gigalomanicx
Gigalomanicx
因为在https://mmdetection3d.readthedocs.io/en/latest/get_started.html 安装pytorch时直接使用了 `condainstall pytorch torchvision -c pytorch` 然而环境的cuda版本低于pytorch2.3.0的11.8，所以自己安装了cpu版（后续检查发现）。 解决方法是对着pytorch官网调整cuda和pytorch版本，然后python projects/BEVFusion/setup.py develop 编译bev_pool_ext
那个pkl是按照2.0生成的吗？ 没改动的话（又是data_preprocessor出问题）应该是数据有问题? 我的BEVFusion复现成功了，有问题可以联系。
Hi @Debrove @afterthat97 , I'm also working on this work. in my opinion, 16 frames cover a fairly long span of around 8 seconds in the NuScenes dataset, which raises...
@leslie27ch I replied to your email and may be able to answer your question.
> @lucianzhong @hopef I am trying to add pcl library to CMakeLists.txt using "find_package (PCL 1.10 REQUIRED)" but when I compile and build the code it shows below error. Can...
It works for me, Thanks !
我也出现了这个问题，不过我使用的是stream_petr_r50_flash_704_bs2_seq_24e.py，flash-attn==0.2.8,训练到5w iters左右梯度爆炸了
我这边可以用test.py设置--task和out_path-可视化img_view，但是bev view 的lidar_det出不来。 
> You can use [this file](https://github.com/DFLyan/notes_of_experiments/blob/master/mmdetection/backup_nusc_explorer.py) or [office file with dataset.show](https://github.com/HuangJunJie2017/BEVDet/blob/dev3.0/tools/misc/visualize_results.py) How to generate the specific pkl？I can only generate result_nucs.json. Thanks for the reply. 哦 我看到你的md了 谢了
@jwli6 你这个能出那个手动标lidar2camera的窗口吗？我这边windows上运行docker desktop 然后在vscode附加跑容器，老是提醒我没装X11 server 也不知道怎么装。