ValueError: cannot convert float NaN to integer

[mtli77]

Hi @jeasinema During the training process, the program throws the following errors， Could you please help me to solve it?

train: 3053/74810 @ epoch:0/10 loss: 0.9977248311042786 reg_loss: 0.6558214426040649 cls_loss: 0.3419033885002136 default train ['aug_003526_1_974'] train: 3054/74810 @ epoch:0/10 loss: 0.325385183095932 reg_loss: 0.0 cls_loss: 0.325385183095932 default train ['aug_004331_3_0_9545'] train: 3055/74810 @ epoch:0/10 loss: 0.7338542938232422 reg_loss: 0.2715231478214264 cls_loss: 0.4623311758041382 default train ['aug_000185_3_1_0412'] train: 3056/74810 @ epoch:0/10 loss: 0.6214444637298584 reg_loss: 0.3237478733062744 cls_loss: 0.297696590423584 default train ['aug_002742_3_0_9607'] train: 3057/74810 @ epoch:0/10 loss: 1.2330245971679688 reg_loss: 0.9148601293563843 cls_loss: 0.31816452741622925 default train ['aug_001406_3_0_9987'] train: 3058/74810 @ epoch:0/10 loss: 0.31853052973747253 reg_loss: 0.0 cls_loss: 0.31853052973747253 default train ['aug_004598_3_0_9636'] train: 3059/74810 @ epoch:0/10 loss: 1.5661653280258179 reg_loss: 1.1667592525482178 cls_loss: 0.3994060456752777 default train ['aug_002038_2_0_3833'] train: 3060/74810 @ epoch:0/10 loss: 2.173323631286621 reg_loss: 1.640864372253418 cls_loss: 0.5324591398239136 default predict ['002727'] /disk/VoxelNet-tensorflow/utils/utils.py:87: RuntimeWarning: invalid value encountered in matmul points = np.matmul(np.array(cfg.MATRIX_T_VELO_2_CAM), points) Traceback (most recent call last): File "train.py", line 146, in tf.app.run(main) File "/home/ubuntu/.conda/envs/tia36/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 125, in run _sys.exit(main(argv)) File "train.py", line 122, in main sess, valid_loader.load(), summary=True) File "/disk/VoxelNet-tensorflow/model/model.py", line 320, in predict_step batch_gt_boxes3d[0]) File "/disk/VoxelNet-tensorflow/utils/utils.py", line 343, in draw_lidar_box3d_on_image projections = lidar_box3d_to_camera_box(boxes3d, cal_projection=True) File "/disk/VoxelNet-tensorflow/utils/utils.py", line 302, in lidar_box3d_to_camera_box minx = int(np.min(points[:, 0])) ValueError: cannot convert float NaN to integer