search

matterport / Mask_RCNN

Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow
Other
24.55k stars 11.69k forks source link

train my own dataset can't detect the object rightly(just show a roi) ，please help me！！！ #2023

Open yangtiming opened 4 years ago

yangtiming commented 4 years ago

I do a project ,about crystal detection, I need use mask rcnn to detect the crystal in the picture ,and need mask them in the picture .after label it ,and train it , and then, test it .

I label it like this ：

![Uploading D732485E186775EC8FDDFD262C0F983D.PNG…]()

and get the labeled picture like this ：

![Uploading 927753C241F4FD4360C0FD4C649B509D.PNG…]()

then，trian them，about 21 pictures。

this is train log:

`nohup: ignoring input 2020-03-01 15:42:12.932951: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.1 instructions, but these are available on your machine and could speed up CPU computations. 2020-03-01 15:42:12.933034: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use SSE4.2 instructions, but these are available on your machine and could speed up CPU computations. 2020-03-01 15:42:12.933044: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX instructions, but these are available on your machine and could speed up CPU computations. 2020-03-01 15:42:12.933052: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX2 instructions, but these are available on your machine and could speed up CPU computations. 2020-03-01 15:42:12.933060: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use AVX512F instructions, but these are available on your machine and could speed up CPU computations. 2020-03-01 15:42:12.933068: W tensorflow/core/platform/cpu_feature_guard.cc:45] The TensorFlow library wasn't compiled to use FMA instructions, but these are available on your machine and could speed up CPU computations.

Configurations: BACKBONE resnet50 BACKBONE_STRIDES [4, 8, 16, 32, 64] BATCH_SIZE 1 BBOX_STD_DEV [0.1 0.1 0.2 0.2] COMPUTE_BACKBONE_SHAPE None DETECTION_MAX_INSTANCES 100 DETECTION_MIN_CONFIDENCE 0.75 DETECTION_NMS_THRESHOLD 0.3 FPN_CLASSIF_FC_LAYERS_SIZE 1024 GPU_COUNT 1 GRADIENT_CLIP_NORM 5.0 IMAGES_PER_GPU 1 IMAGE_MAX_DIM 640 IMAGE_META_SIZE 14 IMAGE_MIN_DIM 512 IMAGE_MIN_SCALE 0 IMAGE_RESIZE_MODE square IMAGE_SHAPE [640 640 3] LEARNING_MOMENTUM 0.9 LEARNING_RATE 0.001 LOSS_WEIGHTS {'rpn_class_loss': 1.0, 'rpn_bbox_loss': 1.0, 'mrcnn_class_loss': 1.0, 'mrcnn_bbox_loss': 1.0, 'mrcnn_mask_loss': 1.0} MASK_POOL_SIZE 14 MASK_SHAPE [28, 28] MAX_GT_INSTANCES 30 MEAN_PIXEL [123.7 116.8 103.9] MINI_MASK_SHAPE (56, 56) NAME shapes NUM_CLASSES 2 POOL_SIZE 7 POST_NMS_ROIS_INFERENCE 1000 POST_NMS_ROIS_TRAINING 2000 ROI_POSITIVE_RATIO 0.33 RPN_ANCHOR_RATIOS [0.5, 1, 2] RPN_ANCHOR_SCALES (48, 96, 192, 384, 768) RPN_ANCHOR_STRIDE 1 RPN_BBOX_STD_DEV [0.1 0.1 0.2 0.2] RPN_NMS_THRESHOLD 0.7 RPN_TRAIN_ANCHORS_PER_IMAGE 256 STEPS_PER_EPOCH 100 TOP_DOWN_PYRAMID_SIZE 256 TRAIN_BN False TRAIN_ROIS_PER_IMAGE 100 USE_MINI_MASK True USE_RPN_ROIS True VALIDATION_STEPS 50 WEIGHT_DECAY 0.0001

21 21_my_data/labelme_json/050_json/img.png 21_my_data/labelme_json/028_json/img.png 21_my_data/labelme_json/06_json/img.png 21_my_data/labelme_json/012_json/img.png 21_my_data/labelme_json/058_json/img.png 21_my_data/labelme_json/039_json/img.png 21_my_data/labelme_json/033_json/img.png 21_my_data/labelme_json/010_json/img.png 21_my_data/labelme_json/017_json/img.png 21_my_data/labelme_json/08_json/img.png 21_my_data/labelme_json/01_json/img.png 21_my_data/labelme_json/07_json/img.png 21_my_data/labelme_json/034_json/img.png 21_my_data/labelme_json/042_json/img.png 21_my_data/labelme_json/016_json/img.png 21_my_data/labelme_json/066_json/img.png 21_my_data/labelme_json/00_json/img.png 21_my_data/labelme_json/011_json/img.png 21_my_data/labelme_json/021_json/img.png 21_my_data/labelme_json/022_json/img.png 21_my_data/labelme_json/043_json/img.png dataset_train--> [ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20] 21_my_data/labelme_json/050_json/img.png 21_my_data/labelme_json/028_json/img.png 21_my_data/labelme_json/06_json/img.png 21_my_data/labelme_json/012_json/img.png 21_my_data/labelme_json/058_json/img.png 21_my_data/labelme_json/039_json/img.png 21_my_data/labelme_json/033_json/img.png 21_my_data/labelme_json/010_json/img.png dataset_val--> [0 1 2 3 4 5 6 7]

Starting at epoch 0. LR=0.001

Checkpoint Path: /home/jovyan/mnt/mask_rcnn_test1111/logs/shapes20200301T1542/mask_rcnnshapes{epoch:04d}.h5 Selecting layers to train fpn_c5p5 (Conv2D) fpn_c4p4 (Conv2D) fpn_c3p3 (Conv2D) fpn_c2p2 (Conv2D) fpn_p5 (Conv2D) fpn_p2 (Conv2D) fpn_p3 (Conv2D) fpn_p4 (Conv2D) In model: rpn_model rpn_conv_shared (Conv2D) rpn_class_raw (Conv2D) rpn_bbox_pred (Conv2D) mrcnn_mask_conv1 (TimeDistributed) mrcnn_mask_bn1 (TimeDistributed) mrcnn_mask_conv2 (TimeDistributed) mrcnn_mask_bn2 (TimeDistributed) mrcnn_class_conv1 (TimeDistributed) mrcnn_class_bn1 (TimeDistributed) mrcnn_mask_conv3 (TimeDistributed) mrcnn_mask_bn3 (TimeDistributed) mrcnn_class_conv2 (TimeDistributed) mrcnn_class_bn2 (TimeDistributed) mrcnn_mask_conv4 (TimeDistributed) mrcnn_mask_bn4 (TimeDistributed) mrcnn_bbox_fc (TimeDistributed) mrcnn_mask_deconv (TimeDistributed) mrcnn_class_logits (TimeDistributed) mrcnn_mask (TimeDistributed) Using TensorFlow backend. /home/jovyan/mnt/.local/lib/python3.6/site-packages/tensorflow/python/ops/gradients_impl.py:95: UserWarning: Converting sparse IndexedSlices to a dense Tensor of unknown shape. This may consume a large amount of memory. "Converting sparse IndexedSlices to a dense Tensor of unknown shape. " /home/jovyan/mnt/.local/lib/python3.6/site-packages/keras/engine/training.py:1987: UserWarning: Using a generator with use_multiprocessing=True and multiple workers may duplicate your data. Please consider using thekeras.utils.Sequence class. UserWarning('Using a generator withuse_multiprocessing=True`' Epoch 1/10

1/100 [..............................] - ETA: 2925s - loss: 5.8216 - rpn_class_loss: 0.0175 - rpn_bbox_loss: 2.8047 - mrcnn_class_loss: 1.3067 - mrcnn_bbox_loss: 0.9313 - mrcnn_mask_loss: 0.7613 2/100 [..............................] - ETA: 2160s - loss: 4.1900 - rpn_class_loss: 0.0137 - rpn_bbox_loss: 1.4549 - mrcnn_class_loss: 1.1134 - mrcnn_bbox_loss: 0.8670 - mrcnn_mask_loss: 0.7410 3/100 [..............................] - ETA: 1819s - loss: 3.6661 - rpn_class_loss: 0.0128 - rpn_bbox_loss: 1.1970 - mrcnn_class_loss: 0.8987 - mrcnn_bbox_loss: 0.8426 - mrcnn_mask_loss: 0.7150 4/100 [>.............................] - ETA: 1647s - loss: 3.4872 - rpn_class_loss: 0.0145 - rpn_bbox_loss: 1.2205 - mrcnn_class_loss: 0.7547 - mrcnn_bbox_loss: 0.8013 - mrcnn_mask_loss: 0.6962 5/100 [>.............................] - ETA: 1548s - loss: 3.0965 - rpn_class_loss: 0.0129 - rpn_bbox_loss: 1.0096 - mrcnn_class_loss: 0.6511 - mrcnn_bbox_loss: 0.7426 - mrcnn_mask_loss: 0.6803 6/100 [>.............................] - ETA: 1478s - loss: 3.1044 - rpn_class_loss: 0.0133 - rpn_bbox_loss: 1.2186 - mrcnn_class_loss: 0.5529 - mrcnn_bbox_loss: 0.6834 - mrcnn_mask_loss: 0.6362 7/100 [=>............................] - ETA: 1423s - loss: 2.8286 - rpn_class_loss: 0.0127 - rpn_bbox_loss: 1.0768 - mrcnn_class_loss: 0.4822 - mrcnn_bbox_loss: 0.6694 - mrcnn_mask_loss: 0.5874 8/100 [=>............................] - ETA: 1374s - loss: 2.8182 - rpn_class_loss: 0.0129 - rpn_bbox_loss: 1.1849 - mrcnn_class_loss: 0.4270 - mrcnn_bbox_loss: 0.6498 - mrcnn_mask_loss: 0.5437 9/100 [=>............................] - ETA: 1339s - loss: 2.7662 - rpn_class_loss: 0.0131 - rpn_bbox_loss: 1.2300 - mrcnn_class_loss: 0.3834 - mrcnn_bbox_loss: 0.6355 - mrcnn_mask_loss: 0.5042 10/100 [==>...........................] - ETA: 1300s - loss: 2.6728 - rpn_class_loss: 0.0209 - rpn_bbox_loss: 1.1695 - mrcnn_class_loss: 0.3460 - mrcnn_bbox_loss: 0.6707 - mrcnn_mask_loss: 0.4656 11/100 [==>...........................] - ETA: 1273s - loss: 2.5071 - rpn_class_loss: 0.0198 - rpn_bbox_loss: 1.0733 - mrcnn_class_loss: 0.3188 - mrcnn_bbox_loss: 0.6541 - mrcnn_mask_loss: 0.4410 12/100 [==>...........................] - ETA: 1247s - loss: 2.3475 - rpn_class_loss: 0.0189 - rpn_bbox_loss: 0.9933 - mrcnn_class_loss: 0.2958 - mrcnn_bbox_loss: 0.6284 - mrcnn_mask_loss: 0.4111 13/100 [==>...........................] - ETA: 1222s - loss: 2.2893 - rpn_class_loss: 0.0184 - rpn_bbox_loss: 0.9776 - mrcnn_class_loss: 0.2794 - mrcnn_bbox_loss: 0.6233 - mrcnn_mask_loss: 0.3905 14/100 [===>..........................] - ETA: 1200s - loss: 2.2357 - rpn_class_loss: 0.0187 - rpn_bbox_loss: 0.9607 - mrcnn_class_loss: 0.2618 - mrcnn_bbox_loss: 0.6256 - mrcnn_mask_loss: 0.3689 15/100 [===>..........................] - ETA: 1177s - loss: 2.1826 - rpn_class_loss: 0.0220 - rpn_bbox_loss: 0.9240 - mrcnn_class_loss: 0.2477 - mrcnn_bbox_loss: 0.6395 - mrcnn_mask_loss: 0.3493 16/100 [===>..........................] - ETA: 1154s - loss: 2.1470 - rpn_class_loss: 0.0232 - rpn_bbox_loss: 0.9071 - mrcnn_class_loss: 0.2339 - mrcnn_bbox_loss: 0.6506 - mrcnn_mask_loss: 0.3322 17/100 [====>.........................] - ETA: 1134s - loss: 2.1039 - rpn_class_loss: 0.0223 - rpn_bbox_loss: 0.8772 - mrcnn_class_loss: 0.2237 - mrcnn_bbox_loss: 0.6599 - mrcnn_mask_loss: 0.3208 18/100 [====>.........................] - ETA: 1114s - loss: 2.0483 - rpn_class_loss: 0.0232 - rpn_bbox_loss: 0.8553 - mrcnn_class_loss: 0.2148 - mrcnn_bbox_loss: 0.6498 - mrcnn_mask_loss: 0.3053 19/100 [====>.........................] - ETA: 1097s - loss: 1.9698 - rpn_class_loss: 0.0222 - rpn_bbox_loss: 0.8150 - mrcnn_class_loss: 0.2036 - mrcnn_bbox_loss: 0.6365 - mrcnn_mask_loss: 0.2926 20/100 [=====>........................] - ETA: 1078s - loss: 1.9802 - rpn_class_loss: 0.0221 - rpn_bbox_loss: 0.8122 - mrcnn_class_loss: 0.1954 - mrcnn_bbox_loss: 0.6679 - mrcnn_mask_loss: 0.2826 21/100 [=====>........................] - ETA: 1061s - loss: 1.9421 - rpn_class_loss: 0.0213 - rpn_bbox_loss: 0.8021 - mrcnn_class_loss: 0.1862 - mrcnn_bbox_loss: 0.6593 - mrcnn_mask_loss: 0.2731 22/100 [=====>........................] - ETA: 1045s - loss: 1.8918 - rpn_class_loss: 0.0205 - rpn_bbox_loss: 0.7667 - mrcnn_class_loss: 0.1778 - mrcnn_bbox_loss: 0.6617 - mrcnn_mask_loss: 0.2650 23/100 [=====>........................] - ETA: 1028s - loss: 1.9125 - rpn_class_loss: 0.0198 - rpn_bbox_loss: 0.8023 - mrcnn_class_loss: 0.1721 - mrcnn_bbox_loss: 0.6604 - mrcnn_mask_loss: 0.2578 24/100 [======>.......................] - ETA: 1012s - loss: 1.8934 - rpn_class_loss: 0.0202 - rpn_bbox_loss: 0.7840 - mrcnn_class_loss: 0.1683 - mrcnn_bbox_loss: 0.6724 - mrcnn_mask_loss: 0.2485 25/100 [======>.......................] - ETA: 997s - loss: 1.8572 - rpn_class_loss: 0.0196 - rpn_bbox_loss: 0.7669 - mrcnn_class_loss: 0.1618 - mrcnn_bbox_loss: 0.6666 - mrcnn_mask_loss: 0.2424  26/100 [======>.......................] - ETA: 982s - loss: 1.8145 - rpn_class_loss: 0.0191 - rpn_bbox_loss: 0.7426 - mrcnn_class_loss: 0.1567 - mrcnn_bbox_loss: 0.6557 - mrcnn_mask_loss: 0.2405 27/100 [=======>......................] - ETA: 966s - loss: 1.7793 - rpn_class_loss: 0.0185 - rpn_bbox_loss: 0.7294 - mrcnn_class_loss: 0.1542 - mrcnn_bbox_loss: 0.6421 - mrcnn_mask_loss: 0.2352 28/100 [=======>......................] - ETA: 951s - loss: 1.7680 - rpn_class_loss: 0.0185 - rpn_bbox_loss: 0.7289 - mrcnn_class_loss: 0.1506 - mrcnn_bbox_loss: 0.6400 - mrcnn_mask_loss: 0.2300 29/100 [=======>......................] - ETA: 937s - loss: 1.7438 - rpn_class_loss: 0.0180 - rpn_bbox_loss: 0.7163 - mrcnn_class_loss: 0.1481 - mrcnn_bbox_loss: 0.6360 - mrcnn_mask_loss: 0.2254 30/100 [========>.....................] - ETA: 922s - loss: 1.7148 - rpn_class_loss: 0.0176 - rpn_bbox_loss: 0.6978 - mrcnn_class_loss: 0.1439 - mrcnn_bbox_loss: 0.6280 - mrcnn_mask_loss: 0.2274 31/100 [========>.....................] - ETA: 908s - loss: 1.6991 - rpn_class_loss: 0.0174 - rpn_bbox_loss: 0.6980 - mrcnn_class_loss: 0.1395 - mrcnn_bbox_loss: 0.6213 - mrcnn_mask_loss: 0.2229 32/100 [========>.....................] - ETA: 893s - loss: 1.6660 - rpn_class_loss: 0.0169 - rpn_bbox_loss: 0.6828 - mrcnn_class_loss: 0.1360 - mrcnn_bbox_loss: 0.6124 - mrcnn_mask_loss: 0.2178 ........... ........... ........... rpn_bbox_loss: 0.0532 - mrcnn_class_loss: 0.0783 - mrcnn_bbox_loss: 0.0656 - mrcnn_mask_loss: 0.0808 28/100 [=======>......................] - ETA: 896s - loss: 0.2770 - rpn_class_loss: 0.0021 - rpn_bbox_loss: 0.0522 - mrcnn_class_loss: 0.0779 - mrcnn_bbox_loss: 0.0645 - mrcnn_mask_loss: 0.0803 29/100 [=======>......................] - ETA: 882s - loss: 0.2752 - rpn_class_loss: 0.0020 - rpn_bbox_loss: 0.0505 - mrcnn_class_loss: 0.0775 - mrcnn_bbox_loss: 0.0639 - mrcnn_mask_loss: 0.0813 30/100 [========>.....................] - ETA: 868s - loss: 0.2737 - rpn_class_loss: 0.0020 - rpn_bbox_loss: 0.0494 - mrcnn_class_loss: 0.0772 - mrcnn_bbox_loss: 0.0627 - mrcnn_mask_loss: 0.0824 31/100 [========>.....................] - ETA: 855s - loss: 0.2718 - rpn_class_loss: 0.0020 - rpn_bbox_loss: 0.0480 - mrcnn_class_loss: 0.0773 - mrcnn_bbox_loss: 0.0616 - mrcnn_mask_loss: 0.0829 32/100 [========>.....................] - ETA: 843s - loss: 0.2819 - rpn_class_loss: 0.0026 - rpn_bbox_loss: 0.0535 - mrcnn_class_loss: 0.0785 - mrcnn_bbox_loss: 0.0652 - mrcnn_mask_loss: 0.0821 33/100 [========>.....................] - ETA: 831s - loss: 0.2875 - rpn_class_loss: 0.0030 - rpn_bbox_loss: 0.0582 - mrcnn_class_loss: 0.0783 - mrcnn_bbox_loss: 0.0668 - mrcnn_mask_loss: 0.0813 34/100 [=========>....................] - ETA: 818s - loss: 0.2854 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0583 - mrcnn_class_loss: 0.0782 - mrcnn_bbox_loss: 0.0659 - mrcnn_mask_loss: 0.0801 35/100 [=========>....................] - ETA: 806s - loss: 0.2863 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0608 - mrcnn_class_loss: 0.0783 - mrcnn_bbox_loss: 0.0655 - mrcnn_mask_loss: 0.0789 36/100 [=========>....................] - ETA: 794s - loss: 0.2890 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0634 - mrcnn_class_loss: 0.0780 - mrcnn_bbox_loss: 0.0663 - mrcnn_mask_loss: 0.0785 37/100 [==========>...................] - ETA: 781s - loss: 0.2863 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0624 - mrcnn_class_loss: 0.0774 - mrcnn_bbox_loss: 0.0662 - mrcnn_mask_loss: 0.0776 38/100 [==========>...................] - ETA: 769s - loss: 0.2871 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0631 - mrcnn_class_loss: 0.0767 - mrcnn_bbox_loss: 0.0673 - mrcnn_mask_loss: 0.0774 39/100 [==========>...................] - ETA: 756s - loss: 0.2908 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0655 - mrcnn_class_loss: 0.0772 - mrcnn_bbox_loss: 0.0687 - mrcnn_mask_loss: 0.0765 40/100 [===========>..................] - ETA: 744s - loss: 0.2873 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0640 - mrcnn_class_loss: 0.0765 - mrcnn_bbox_loss: 0.0685 - mrcnn_mask_loss: 0.0756 41/100 [===========>..................] - ETA: 731s - loss: 0.2946 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0693 - mrcnn_class_loss: 0.0779 - mrcnn_bbox_loss: 0.0694 - mrcnn_mask_loss: 0.0751 42/100 [===========>..................] - ETA: 719s - loss: 0.2933 - rpn_class_loss: 0.0030 - rpn_bbox_loss: 0.0678 - mrcnn_class_loss: 0.0776 - mrcnn_bbox_loss: 0.0704 - mrcnn_mask_loss: 0.0746 43/100 [===========>..................] - ETA: 707s - loss: 0.2922 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0663 - mrcnn_class_loss: 0.0771 - mrcnn_bbox_loss: 0.0706 - mrcnn_mask_loss: 0.0752 44/100 [============>.................] - ETA: 694s - loss: 0.2898 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0654 - mrcnn_class_loss: 0.0767 - mrcnn_bbox_loss: 0.0705 - mrcnn_mask_loss: 0.0744 45/100 [============>.................] - ETA: 682s - loss: 0.2885 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0643 - mrcnn_class_loss: 0.0769 - mrcnn_bbox_loss: 0.0703 - mrcnn_mask_loss: 0.0738 46/100 [============>.................] - ETA: 669s - loss: 0.2868 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0637 - mrcnn_class_loss: 0.0761 - mrcnn_bbox_loss: 0.0694 - mrcnn_mask_loss: 0.0745 47/100 [=============>................] - ETA: 658s - loss: 0.2909 - rpn_class_loss: 0.0032 - rpn_bbox_loss: 0.0651 - mrcnn_class_loss: 0.0768 - mrcnn_bbox_loss: 0.0719 - mrcnn_mask_loss: 0.0739 48/100 [=============>................] - ETA: 645s - loss: 0.2898 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0641 - mrcnn_class_loss: 0.0765 - mrcnn_bbox_loss: 0.0721 - mrcnn_mask_loss: 0.0739 49/100 [=============>................] - ETA: 633s - loss: 0.2890 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0632 - mrcnn_class_loss: 0.0769 - mrcnn_bbox_loss: 0.0723 - mrcnn_mask_loss: 0.0735 50/100 [==============>...............] - ETA: 620s - loss: 0.2949 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0653 - mrcnn_class_loss: 0.0799 - mrcnn_bbox_loss: 0.0742 - mrcnn_mask_loss: 0.0724 51/100 [==============>...............] - ETA: 608s - loss: 0.2932 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0645 - mrcnn_class_loss: 0.0793 - mrcnn_bbox_loss: 0.0735 - mrcnn_mask_loss: 0.0729 52/100 [==============>...............] - ETA: 595s - loss: 0.2907 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0635 - mrcnn_class_loss: 0.0788 - mrcnn_bbox_loss: 0.0732 - mrcnn_mask_loss: 0.0721 53/100 [==============>...............] - ETA: 583s - loss: 0.2869 - rpn_class_loss: 0.0030 - rpn_bbox_loss: 0.0623 - mrcnn_class_loss: 0.0780 - mrcnn_bbox_loss: 0.0721 - mrcnn_mask_loss: 0.0715 54/100 [===============>..............] - ETA: 571s - loss: 0.2877 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0630 - mrcnn_class_loss: 0.0790 - mrcnn_bbox_loss: 0.0718 - mrcnn_mask_loss: 0.0709 55/100 [===============>..............] - ETA: 559s - loss: 0.2873 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0622 - mrcnn_class_loss: 0.0785 - mrcnn_bbox_loss: 0.0713 - mrcnn_mask_loss: 0.0722 56/100 [===============>..............] - ETA: 547s - loss: 0.2853 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0612 - mrcnn_class_loss: 0.0780 - mrcnn_bbox_loss: 0.0709 - mrcnn_mask_loss: 0.0720 57/100 [================>.............] - ETA: 534s - loss: 0.2832 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0602 - mrcnn_class_loss: 0.0773 - mrcnn_bbox_loss: 0.0700 - mrcnn_mask_loss: 0.0725 58/100 [================>.............] - ETA: 522s - loss: 0.2815 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0595 - mrcnn_class_loss: 0.0769 - mrcnn_bbox_loss: 0.0698 - mrcnn_mask_loss: 0.0722 59/100 [================>.............] - ETA: 510s - loss: 0.2819 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0586 - mrcnn_class_loss: 0.0767 - mrcnn_bbox_loss: 0.0708 - mrcnn_mask_loss: 0.0726 60/100 [=================>............] - ETA: 498s - loss: 0.2801 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0577 - mrcnn_class_loss: 0.0760 - mrcnn_bbox_loss: 0.0706 - mrcnn_mask_loss: 0.0727 61/100 [=================>............] - ETA: 485s - loss: 0.2795 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0572 - mrcnn_class_loss: 0.0753 - mrcnn_bbox_loss: 0.0706 - mrcnn_mask_loss: 0.0732 62/100 [=================>............] - ETA: 473s - loss: 0.2779 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0565 - mrcnn_class_loss: 0.0745 - mrcnn_bbox_loss: 0.0702 - mrcnn_mask_loss: 0.0736 63/100 [=================>............] - ETA: 460s - loss: 0.2771 - rpn_class_loss: 0.0031 - rpn_bbox_loss: 0.0556 - mrcnn_class_loss: 0.0740 - mrcnn_bbox_loss: 0.0698 - mrcnn_mask_loss: 0.0747 64/100 [==================>...........] - ETA: 448s - loss: 0.2755 - rpn_class_loss: 0.0030 - rpn_bbox_loss: 0.0549 - mrcnn_class_loss: 0.0734 - mrcnn_bbox_loss: 0.0690 - mrcnn_mask_loss: 0.0751 65/100 [==================>...........] - ETA: 435s - loss: 0.2753 - rpn_class_loss: 0.0030 - rpn_bbox_loss: 0.0541 - mrcnn_class_loss: 0.0740 - mrcnn_bbox_loss: 0.0688 - mrcnn_mask_loss: 0.0754 66/100 [==================>...........] - ETA: 423s - loss: 0.2746 - rpn_class_loss: 0.0030 - rpn_bbox_loss: 0.0534 - mrcnn_class_loss: 0.0736 - mrcnn_bbox_loss: 0.0684 - mrcnn_mask_loss: 0.0763 67/100 [===================>..........] - ETA: 411s - loss: 0.2732 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0530 - mrcnn_class_loss: 0.0734 - mrcnn_bbox_loss: 0.0683 - mrcnn_mask_loss: 0.0757 68/100 [===================>..........] - ETA: 398s - loss: 0.2716 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0523 - mrcnn_class_loss: 0.0731 - mrcnn_bbox_loss: 0.0676 - mrcnn_mask_loss: 0.0756 69/100 [===================>..........] - ETA: 386s - loss: 0.2698 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0516 - mrcnn_class_loss: 0.0729 - mrcnn_bbox_loss: 0.0669 - mrcnn_mask_loss: 0.0755 70/100 [====================>.........] - ETA: 373s - loss: 0.2678 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0510 - mrcnn_class_loss: 0.0728 - mrcnn_bbox_loss: 0.0661 - mrcnn_mask_loss: 0.0751 71/100 [====================>.........] - ETA: 361s - loss: 0.2666 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0503 - mrcnn_class_loss: 0.0730 - mrcnn_bbox_loss: 0.0660 - mrcnn_mask_loss: 0.0745 72/100 [====================>.........] - ETA: 348s - loss: 0.2672 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0512 - mrcnn_class_loss: 0.0729 - mrcnn_bbox_loss: 0.0661 - mrcnn_mask_loss: 0.0742 73/100 [====================>.........] - ETA: 336s - loss: 0.2657 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0506 - mrcnn_class_loss: 0.0728 - mrcnn_bbox_loss: 0.0654 - mrcnn_mask_loss: 0.0740 74/100 [=====================>........] - ETA: 324s - loss: 0.2655 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0500 - mrcnn_class_loss: 0.0730 - mrcnn_bbox_loss: 0.0652 - mrcnn_mask_loss: 0.0743 75/100 [=====================>........] - ETA: 311s - loss: 0.2642 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0494 - mrcnn_class_loss: 0.0727 - mrcnn_bbox_loss: 0.0650 - mrcnn_mask_loss: 0.0743 76/100 [=====================>........] - ETA: 299s - loss: 0.2624 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0487 - mrcnn_class_loss: 0.0724 - mrcnn_bbox_loss: 0.0647 - mrcnn_mask_loss: 0.0737 77/100 [======================>.......] - ETA: 286s - loss: 0.2616 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0489 - mrcnn_class_loss: 0.0718 - mrcnn_bbox_loss: 0.0648 - mrcnn_mask_loss: 0.0732 78/100 [======================>.......] - ETA: 274s - loss: 0.2606 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0483 - mrcnn_class_loss: 0.0717 - mrcnn_bbox_loss: 0.0647 - mrcnn_mask_loss: 0.0731 79/100 [======================>.......] - ETA: 261s - loss: 0.2591 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0477 - mrcnn_class_loss: 0.0713 - mrcnn_bbox_loss: 0.0642 - mrcnn_mask_loss: 0.0731 80/100 [=======================>......] - ETA: 249s - loss: 0.2637 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0493 - mrcnn_class_loss: 0.0732 - mrcnn_bbox_loss: 0.0660 - mrcnn_mask_loss: 0.0724 81/100 [=======================>......] - ETA: 236s - loss: 0.2649 - rpn_class_loss: 0.0029 - rpn_bbox_loss: 0.0492 - mrcnn_class_loss: 0.0729 - mrcnn_bbox_loss: 0.0667 - mrcnn_mask_loss: 0.0733 82/100 [=======================>......] - ETA: 224s - loss: 0.2639 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0488 - mrcnn_class_loss: 0.0724 - mrcnn_bbox_loss: 0.0667 - mrcnn_mask_loss: 0.0733 83/100 [=======================>......] - ETA: 211s - loss: 0.2640 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0484 - mrcnn_class_loss: 0.0718 - mrcnn_bbox_loss: 0.0674 - mrcnn_mask_loss: 0.0736 84/100 [========================>.....] - ETA: 199s - loss: 0.2634 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0480 - mrcnn_class_loss: 0.0713 - mrcnn_bbox_loss: 0.0674 - mrcnn_mask_loss: 0.0739 85/100 [========================>.....] - ETA: 186s - loss: 0.2621 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0475 - mrcnn_class_loss: 0.0708 - mrcnn_bbox_loss: 0.0668 - mrcnn_mask_loss: 0.0742 86/100 [========================>.....] - ETA: 174s - loss: 0.2603 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0470 - mrcnn_class_loss: 0.0706 - mrcnn_bbox_loss: 0.0663 - mrcnn_mask_loss: 0.0737 87/100 [=========================>....] - ETA: 161s - loss: 0.2597 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0465 - mrcnn_class_loss: 0.0706 - mrcnn_bbox_loss: 0.0663 - mrcnn_mask_loss: 0.0736 88/100 [=========================>....] - ETA: 149s - loss: 0.2584 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0462 - mrcnn_class_loss: 0.0703 - mrcnn_bbox_loss: 0.0661 - mrcnn_mask_loss: 0.0731 89/100 [=========================>....] - ETA: 137s - loss: 0.2588 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0458 - mrcnn_class_loss: 0.0702 - mrcnn_bbox_loss: 0.0670 - mrcnn_mask_loss: 0.0730 90/100 [==========================>...] - ETA: 124s - loss: 0.2572 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0454 - mrcnn_class_loss: 0.0700 - mrcnn_bbox_loss: 0.0665 - mrcnn_mask_loss: 0.0727 91/100 [==========================>...] - ETA: 112s - loss: 0.2611 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0462 - mrcnn_class_loss: 0.0719 - mrcnn_bbox_loss: 0.0682 - mrcnn_mask_loss: 0.0721 92/100 [==========================>...] - ETA: 99s - loss: 0.2611 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0465 - mrcnn_class_loss: 0.0721 - mrcnn_bbox_loss: 0.0679 - mrcnn_mask_loss: 0.0719  93/100 [==========================>...] - ETA: 87s - loss: 0.2599 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0462 - mrcnn_class_loss: 0.0719 - mrcnn_bbox_loss: 0.0676 - mrcnn_mask_loss: 0.0715 94/100 [===========================>..] - ETA: 74s - loss: 0.2618 - rpn_class_loss: 0.0027 - rpn_bbox_loss: 0.0467 - mrcnn_class_loss: 0.0733 - mrcnn_bbox_loss: 0.0681 - mrcnn_mask_loss: 0.0709 95/100 [===========================>..] - ETA: 62s - loss: 0.2614 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0463 - mrcnn_class_loss: 0.0729 - mrcnn_bbox_loss: 0.0678 - mrcnn_mask_loss: 0.0716 96/100 [===========================>..] - ETA: 49s - loss: 0.2617 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0458 - mrcnn_class_loss: 0.0730 - mrcnn_bbox_loss: 0.0689 - mrcnn_mask_loss: 0.0712 97/100 [============================>.] - ETA: 37s - loss: 0.2610 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0453 - mrcnn_class_loss: 0.0726 - mrcnn_bbox_loss: 0.0685 - mrcnn_mask_loss: 0.0719 98/100 [============================>.] - ETA: 24s - loss: 0.2611 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0457 - mrcnn_class_loss: 0.0726 - mrcnn_bbox_loss: 0.0684 - mrcnn_mask_loss: 0.0716 99/100 [============================>.] - ETA: 12s - loss: 0.2605 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0454 - mrcnn_class_loss: 0.0724 - mrcnn_bbox_loss: 0.0684 - mrcnn_mask_loss: 0.0714 100/100 [==============================] - 1416s - loss: 0.2596 - rpn_class_loss: 0.0028 - rpn_bbox_loss: 0.0451 - mrcnn_class_loss: 0.0723 - mrcnn_bbox_loss: 0.0681 - mrcnn_mask_loss: 0.0714 - val_loss: 0.2719 - val_rpn_class_loss: 0.0037 - val_rpn_bbox_loss: 0.0128 - val_mrcnn_class_loss: 0.0786 - val_mrcnn_bbox_loss: 0.1018 - val_mrcnn_mask_loss: 0.0751

Starting at epoch 10. LR=0.0001

Checkpoint Path: /home/jovyan/mnt/mask_rcnn_test1111/logs/shapes20200301T1542/mask_rcnnshapes{epoch:04d}.h5 Selecting layers to train conv1 (Conv2D) bn_conv1 (BatchNorm) res2a_branch2a (Conv2D) bn2a_branch2a (BatchNorm) res2a_branch2b (Conv2D) bn2a_branch2b (BatchNorm) res2a_branch2c (Conv2D) res2a_branch1 (Conv2D) bn2a_branch2c (BatchNorm) bn2a_branch1 (BatchNorm) res2b_branch2a (Conv2D) bn2b_branch2a (BatchNorm) res2b_branch2b (Conv2D) bn2b_branch2b (BatchNorm) res2b_branch2c (Conv2D) bn2b_branch2c (BatchNorm) res2c_branch2a (Conv2D) bn2c_branch2a (BatchNorm) res2c_branch2b (Conv2D) bn2c_branch2b (BatchNorm) res2c_branch2c (Conv2D) bn2c_branch2c (BatchNorm) res3a_branch2a (Conv2D) bn3a_branch2a (BatchNorm) res3a_branch2b (Conv2D) bn3a_branch2b (BatchNorm) res3a_branch2c (Conv2D) res3a_branch1 (Conv2D) bn3a_branch2c (BatchNorm) bn3a_branch1 (BatchNorm) res3b_branch2a (Conv2D) bn3b_branch2a (BatchNorm) res3b_branch2b (Conv2D) bn3b_branch2b (BatchNorm) res3b_branch2c (Conv2D) bn3b_branch2c (BatchNorm) res3c_branch2a (Conv2D) bn3c_branch2a (BatchNorm) res3c_branch2b (Conv2D) bn3c_branch2b (BatchNorm) res3c_branch2c (Conv2D) bn3c_branch2c (BatchNorm) res3d_branch2a (Conv2D) bn3d_branch2a (BatchNorm) res3d_branch2b (Conv2D) bn3d_branch2b (BatchNorm) res3d_branch2c (Conv2D) bn3d_branch2c (BatchNorm) res4a_branch2a (Conv2D) bn4a_branch2a (BatchNorm) res4a_branch2b (Conv2D) bn4a_branch2b (BatchNorm) res4a_branch2c (Conv2D) res4a_branch1 (Conv2D) bn4a_branch2c (BatchNorm) bn4a_branch1 (BatchNorm) res4b_branch2a (Conv2D) bn4b_branch2a (BatchNorm) res4b_branch2b (Conv2D) bn4b_branch2b (BatchNorm) res4b_branch2c (Conv2D) bn4b_branch2c (BatchNorm) res4c_branch2a (Conv2D) bn4c_branch2a (BatchNorm) res4c_branch2b (Conv2D) bn4c_branch2b (BatchNorm) res4c_branch2c (Conv2D) bn4c_branch2c (BatchNorm) res4d_branch2a (Conv2D) bn4d_branch2a (BatchNorm) res4d_branch2b (Conv2D) bn4d_branch2b (BatchNorm) res4d_branch2c (Conv2D) bn4d_branch2c (BatchNorm) res4e_branch2a (Conv2D) bn4e_branch2a (BatchNorm) res4e_branch2b (Conv2D) bn4e_branch2b (BatchNorm) res4e_branch2c (Conv2D) bn4e_branch2c (BatchNorm) res4f_branch2a (Conv2D) bn4f_branch2a (BatchNorm) res4f_branch2b (Conv2D) bn4f_branch2b (BatchNorm) res4f_branch2c (Conv2D) bn4f_branch2c (BatchNorm) res5a_branch2a (Conv2D) bn5a_branch2a (BatchNorm) res5a_branch2b (Conv2D) bn5a_branch2b (BatchNorm) res5a_branch2c (Conv2D) res5a_branch1 (Conv2D) bn5a_branch2c (BatchNorm) bn5a_branch1 (BatchNorm) res5b_branch2a (Conv2D) bn5b_branch2a (BatchNorm) res5b_branch2b (Conv2D) bn5b_branch2b (BatchNorm) res5b_branch2c (Conv2D) bn5b_branch2c (BatchNorm) res5c_branch2a (Conv2D) bn5c_branch2a (BatchNorm) res5c_branch2b (Conv2D) bn5c_branch2b (BatchNorm) res5c_branch2c (Conv2D) bn5c_branch2c (BatchNorm) fpn_c5p5 (Conv2D) fpn_c4p4 (Conv2D) fpn_c3p3 (Conv2D) fpn_c2p2 (Conv2D) fpn_p5 (Conv2D) fpn_p2 (Conv2D) fpn_p3 (Conv2D) fpn_p4 (Conv2D) In model: rpn_model rpn_conv_shared (Conv2D) rpn_class_raw (Conv2D) rpn_bbox_pred (Conv2D) mrcnn_mask_conv1 (TimeDistributed) mrcnn_mask_bn1 (TimeDistributed) mrcnn_mask_conv2 (TimeDistributed) mrcnn_mask_bn2 (TimeDistributed) mrcnn_class_conv1 (TimeDistributed) mrcnn_class_bn1 (TimeDistributed) mrcnn_mask_conv3 (TimeDistributed) mrcnn_mask_bn3 (TimeDistributed) mrcnn_class_conv2 (TimeDistributed) mrcnn_class_bn2 (TimeDistributed) mrcnn_mask_conv4 (TimeDistributed) mrcnn_mask_bn4 (TimeDistributed) mrcnn_bbox_fc (TimeDistributed) mrcnn_mask_deconv (TimeDistributed) mrcnn_class_logits (TimeDistributed) mrcnn_mask (TimeDistributed)

`

but the result let me disappointed。 the result below:

![Uploading DDD.PNG…]()

![Uploading F0A21890F0C77ED6A111DACDA6B0B1AB.PNG…]()

suchiz commented 4 years ago

Hi there, just a few things: