yudefu

home/zzz/Downloads/UniAD-1.0.1/projects/mmdet3d_plugin/datasets/pipelines/loading.py把这个脚本里面第48行注释掉；你这个很明显是两个目录join操作导致本身目录文件存在，结果访问不到

你这种问题好像是安装环境没有搞好，我看很多人在conda环境里面都遇见一些问题，不建议在conda环境安装；你可以在非conda环境进行安装，如果你的cuda版本不匹配可以参考：https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&Distribution=CentOS&target_version=7&target_type=runfile_local进行安装；如果需要安装torch相关的包可以参考：https://download.pytorch.org/whl/torch_stable.html

> > home/zzz/Downloads/UniAD-1.0.1/projects/mmdet3d_plugin/datasets/pipelines/loading.py把这个脚本里面第48行注释掉；你这个很明显是两个目录join操作导致本身目录文件存在，结果访问不到 > > > > 你好，请问下载mini数据集后data文件夹应该是什么结构呢？使用mini集时，data_root t改为空empty在哪一个config files里修改呢？我使用./uniad_create_data.sh时出现了如下的报错： > > > (uniad) cyang@cyang-Precision-3660:~/UniAD/tools$ ./uniad_create_data.sh > > > v1.0-mini /home/cyang/UniAD/data/nuscenes > > > ====== > > > Loading NuScenes...

> @yudefu Hi, both of your installed mmcv and mmdet3d have higher version than what's required in this project. You may need to create a new environment and install the...

> Hi authors, > > Thanks for the great work. It is truly impressive! Regarding Exp.10 in Table 2, I have a few questions about the 'planning-only' experiment. Could you...

------------------ 原始邮件 ------------------ 发件人: "OpenDriveLab/UniAD" ***@***.***>; 发送时间: 2023年7月31日(星期一) 下午2:47 ***@***.***>; ***@***.******@***.***>; 主题: Re: [OpenDriveLab/UniAD] Question Regarding Exp.10 in Table 2 - Training Method for Planning-Only Experiment (Issue #93) 你好，可以加我微信吗？y1193547749 Could we possibly use WeChat...

看一下你的uniad_dist_eval.sh里面的python -m torch.distributed.launch是不是要改成python3 -m torch.distributed.launch；有可能你的conda环境有关系

10个blogs全部下载下来，他们的目录层级是一样的，需要把他们合并到一起，当然最后可能因为一些文件重复会导致总数目对不上，但是不影响训练；提供两个脚本： transfer_samples_data.sh： #!bin/bash path=$1 version_name=$2 mv ${path}/${version_name}/samples/CAM_BACK/* ./data/nuscenes/samples/CAM_BACK mv ${path}/${version_name}/samples/CAM_BACK_LEFT/* ./data/nuscenes/samples/CAM_BACK_LEFT/ mv ${path}/${version_name}/samples/CAM_BACK_RIGHT/* ./data/nuscenes/samples/CAM_BACK_RIGHT/ mv ${path}/${version_name}/samples/CAM_FRONT/* ./data/nuscenes/samples/CAM_FRONT mv ${path}/${version_name}/samples/CAM_FRONT_LEFT/* ./data/nuscenes/samples/CAM_FRONT_LEFT/ mv ${path}/${version_name}/samples/CAM_FRONT_RIGHT/* ./data/nuscenes/samples/CAM_FRONT_RIGHT/ mv ${path}/${version_name}/samples/LIDAR_TOP/* ./data/nuscenes/samples/LIDAR_TOP/ mv ${path}/${version_name}/samples/RADAR_BACK_LEFT/* ./data/nuscenes/samples/RADAR_BACK_LEFT/ mv...

> > 非常感谢！发自我的iPhone------------------ Original ------------------From: yudefu _**@**_._**>Date: Thu,Sep 7,2023 3:29 PMTo: OpenDriveLab/UniAD _**@**_.**_>Cc: stephenunderdog _**@**_._**>, Mention _**@**_.**_>Subject: Re: [OpenDriveLab/UniAD] How to use a full nusences dataset that hasten folders and...

> Same Question. 你好，你跑成功训练stage1或者测试那个例子了吗