电子厂小熊

SUMMARY ======== PARAMETERS * /moveit_simple_grasps_server/base_link: base_link * /moveit_simple_grasps_server/end_effector: gripper * /moveit_simple_grasps_server/gripper/end_effector_name: gripper * /moveit_simple_grasps_server/gripper/end_effector_parent_link: gripper_roll_link * /moveit_simple_grasps_server/gripper/grasp_pose_to_eef: [0.0, 0.0, 0.0] * /moveit_simple_grasps_server/gripper/grasp_pose_to_eef_rotation: [0.0, 0.0, 0.0] * /moveit_simple_grasps_server/gripper/grasp_posture: [1.0, 1.0] *...

Hi @wangcongrobot 我也遇到了同样的问题，系统版本的问题，而且这个package应该是indigo下的

> In my environment, > > ``` > ret, err := IOCounters("vda1") > fmt.Println(ret, err) > ``` > > returns this. > > ``` > map[vda1:{"readCount":68022,"mergedReadCount":5,"writeCount":768588,"mergedWriteCount":686205,"readBytes":1546142720,"writeBytes":14503629824,"readTime":62068,"writeTime":697344,"iopsInProgress":0,"ioTime":329020,"weightedIO":759240,"name":"vda1","serialNumber":""}] > ``` hi, when...

> Change the value after `--gpu` option as 1 in "command_train_v1(or v2).sh" Thanks for your help. But I mean that how can i use the 2 gpus working parallely. "--gpu=1"...

> @liqisa Did you do this successfully? https://github.com/lujiaxuan0520/PointCNN-self-driving-cars It's possible to face overfeat when training PointCNN on KITTI PointCloud ,as the author saying.