MARUIHAN123
MARUIHAN123
我也是这种情况。是由于class_loss随着epoch会变成nan. 我是自己的数据集。请问怎么解决
424/3425 [============================>.] - ETA: 0s - loss: nan - rpn_bbox_loss: 0.4054 - rpn_class_loss: 0.1190 - rcnn_bbox_loss: 0.2831 - rcnn_class_loss: nan - regular_loss: 13.5767 - gt_num: 2.2009 - positive_anchor_num: 27.8823 -...
python3.7. tensorflow1.14.0 keras 2.3.1 我是用的自己的数据集。数据集是按照voc的格式做的。麻烦您帮我看一下。谢谢 WARNING:tensorflow:Model failed to serialize as JSON. Ignoring... ('Not JSON Serializable:', b'\n\x04Mean\x12\x04Mean\x1a\x12rpn_bbox_loss/Mean\x1a\x05Const*\x0f\n\tkeep_dims\x12\x02(\x01*\n\n\x04Tidx\x12\x020\x03*\x07\n\x01T\x12\x020\x01') Epoch 1/80 Traceback (most recent call last): File "/home/multiai4/Ruihan/face_mask_detection/train.py", line 133, in main(arguments) File...
The prediction categories do not correspond, but the accuracy is very high for example, my dataset has file categories 1,2,3 When predicting, the picture in category 1 is predicted to...