jiaxin
> Yes Thank you
@xxa783 Could you tell me the amount of your GPUs for training? I guess this is an ignored factor for reproducing the result.
I think it's a typo too. the layer norm2 norm3 are not utilized in the forword
我也是对这个很有疑惑，逻辑上讲不通的感觉，你想明白了吗
@Liu202209 你调整完后还能复现吗，我调整完后反而不能复现了， Car [email protected], 0.70, 0.70: bbox AP:96.6737, 89.8503, 89.5701 bev AP:90.2488, 88.5818, 88.2505 3d AP:89.5764, 84.6657, 84.8719 aos AP:96.65, 89.78, 89.43 Car [email protected], 0.70, 0.70: bbox AP:98.2332, 95.1310, 94.8786...
@Liu202209 我觉得这个对小目标的影响还是很大的 毕竟最后的spatial_feature对应的物理尺寸是8*0.05=0.4 cyclist 和 pedestrian的anchor尺寸也不过是[ 1.76, 0.6, 1.73 ]和[ 0.8, 0.6, 1.73 ] （顺序为lwh） 我多类别调整前的结果是 Car [email protected], 0.70, 0.70: bbox AP:98.1120, 94.6818, 89.4393 bev AP:90.4446, 88.7608, 88.0496 3d AP:89.7993,...
单类别训练Car确实可以达到文章的精度，多类别感觉达不到了
CLASS_NAMES: ['Car', 'Pedestrian', 'Cyclist'] DATA_CONFIG: _BASE_CONFIG_: cfgs/dataset_configs/kitti_dataset.yaml DATASET: 'KittiDataset' ROT_NUM: 3 USE_VAN: True DATA_SPLIT: { 'train': train, 'test': val } INFO_PATH: { 'train': [kitti_infos_train.pkl], 'test': [kitti_infos_val.pkl], } DATA_AUGMENTOR: DISABLE_AUG_LIST: ['placeholder']...
@csl1994 @Liu202209 你们用八卡训练过吗，复现出来了吗？我一直在想是不是训练卡数量的问题