我一直在使用tensorflow object_detection模型和ssd_inception_v2_coco。我设法成功地训练了模型,但是在评估过程中不断遇到此警告。这就是我不断得到的东西。
C:\Users\Keshav\tensorflo>python object_detection/eval.py --logtostderr --pipeline_config_path=data\ssd_inception_v2_coco_sim.config --checkpoint_dir=models\train --eval_dir=eval
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
INFO:tensorflow:depth of additional conv before box predictor: 0
2018-10-20 21:06:20.397576: I T:\src\github\tensorflow\tensorflow\core\platform\cpu_feature_guard.cc:140] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
INFO:tensorflow:Restoring parameters from models\train\model.ckpt-8
INFO:tensorflow:Restoring parameters from models\train\model.ckpt-8
WARNING:root:image 0 does not have groundtruth difficult flag specified
WARNING:root:The following classes have no ground truth examples: 4
C:\Users\Keshav\tensorflo\models\research\object_detection\utils\metrics.py:144: RuntimeWarning: invalid value encountered in true_divide
num_images_correctly_detected_per_class / num_gt_imgs_per_class)
RuntimeWarning之后,它会在长时间挂起后退出并返回到提示,并将事件文件打印到eval目录中。当我在TensorBoard上可视化数据时,一切都为0,但是我认为这可能是因为训练步骤数量极少。我需要担心警告吗?