Affected Cell model

.gitignore

@@ -19,6 +19,8 @@ Result_yolo3_fault_4/result_otros/
 Result_Ortofoto/
 Result_Complete/
 Result_Complete_Example/
+Problem_weights_saved.ipynb
 Train&Test_H/
+
 *.h5
 *.tif

README.md

@@ -190,8 +190,7 @@ All of weights of this trained model grab from [Drive_Weights](https://drive.goo
 |   YOLO3 Panel  |      [weight](https://drive.google.com/open?id=14zgtgDJv3KTvhRC-VOz6sqsGPC_bdrL1)      | [config](config_full_yolo_panel_infer.json) |
 |  YOLO3 Soiling |      [weight](https://drive.google.com/open?id=1YLgkn1wL5xAGOpwd2gzdfsJVGYPzszn-)      | [config](config_full_yolo_fault_1_infer.json) |
 |   YOLO3 Diode  |      [weight](https://drive.google.com/open?id=1VUtrK9JVTbzBw5dX7_dgLTMToFHbAJl1)      | [config](config_full_yolo_fault_4_infer.json) |
-
-|   YOLO3 Affected Cell  |      [weight_lack](...)      | [config](config_full_yolo_fault_2_infer.json) |
+|   YOLO3 Affected Cell    |      [weight(https://drive.google.com/open?id=1ngyCzw7xF0N5oZnF29EIS5LOl1PFkRRM)      | [config](config_full_yolo_fault_2_infer.json) |
 
 ## Panel Detector
 ### SDD7
@@ -221,13 +220,21 @@ On folder [Result yolo3 fault 1](Result_yolo3_fault_1/) show [history train](Res
 
 ![](Result_yolo3_fault_1/result_yolo3_fault_1/Mision_11_DJI_0011.jpg)
 
+## Diode Affected Cell Detector
+### YOLO3
+On folder [Result yolo3 fault 2](Result_yolo3_fault_2/) show [history train](Result_yolo3_fault_2/yolo3_full_yolo.output), weight and result of this model (mAP 71.93%).
+
+![](Result_yolo3_fault_2/result_yolo3_fault_2/Mision%2010_DJI_0093.jpg)
 
 ## Diode Fault Detector
 ### YOLO3
-On folder [Result yolo3 fault 4](Result_yolo3_fault_4/) show [history train](Result_yolo3_fault_4/yolo3_full_yolo.output), weight and result of this model (mAP 73.02%).
+On folder [Result yolo3 fault 4](Result_yolo3_fault_4/) show [history train](Result_yolo3_fault_4/yolo3_full_yolo.output), weight and result of this model (mAP 66.22%).
 
 ![](Result_yolo3_fault_4/result_yolo3_fault_4/Mision%2041_DJI_0044.jpg)
 
+
+
+
 ## Panel Disconnect Detector
 ### YOLO3
 To use the detector we must only use 'panel_yolo3_disconnect.py' with the previously established form, that is:
@@ -257,3 +264,4 @@ Before sending your pull requests, make sure you followed this list.
 # Example to use trained model
 
 In ['Example_Prediction'](Example_prediction.ipynb) this is the example of how to implement an already trained model, it can be modified to change the model you have to use and the image in which you want to detect faults.
+In ['Example_Prediction_AllInOne'](Example Detection AllInOne.ipynb) this is the example of how implement all trained model, you can use this code for predict a folder of images and have a output image with detection boxes.

Result_yolo3_fault_2/.gitignore

@@ -0,0 +1 @@
+log_experimento_fault_gpu_2/

Result_yolo3_fault_2/result_yolo3_fault_2/time.txt

@@ -0,0 +1 @@
+Tiempo promedio:1.493388593196869

Result_yolo3_fault_2/yolo3_full_yolo.err

@@ -0,0 +1,32 @@
+2020-06-15 17:18:13.491693: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer.so.6'; dlerror: libnvinfer.so.6: cannot open shared object file: No such file or directory
+2020-06-15 17:18:13.491844: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer_plugin.so.6'; dlerror: libnvinfer_plugin.so.6: cannot open shared object file: No such file or directory
+2020-06-15 17:18:13.491866: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:30] Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
+2020-06-15 17:18:15.074534: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory
+2020-06-15 17:18:15.074590: E tensorflow/stream_executor/cuda/cuda_driver.cc:351] failed call to cuInit: UNKNOWN ERROR (303)
+2020-06-15 17:18:15.074642: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (dlsaavedra-X406UAR): /proc/driver/nvidia/version does not exist
+2020-06-15 17:18:15.074938: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
+2020-06-15 17:18:15.108036: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 1800000000 Hz
+2020-06-15 17:18:15.110399: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x55c1bd39caf0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
+2020-06-15 17:18:15.110513: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
+WARNING:tensorflow:AutoGraph could not transform <bound method YoloLayer.call of <yolo.YoloLayer object at 0x7facc40764d0>> and will run it as-is.
+Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.
+Cause: unexpected indent (<unknown>, line 144)
+WARNING:tensorflow:AutoGraph could not transform <bound method YoloLayer.call of <yolo.YoloLayer object at 0x7faca44dff90>> and will run it as-is.
+Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.
+Cause: unexpected indent (<unknown>, line 144)
+WARNING:tensorflow:AutoGraph could not transform <bound method YoloLayer.call of <yolo.YoloLayer object at 0x7faca42a9a10>> and will run it as-is.
+Please report this to the TensorFlow team. When filing the bug, set the verbosity to 10 (on Linux, `export AUTOGRAPH_VERBOSITY=10`) and attach the full output.
+Cause: unexpected indent (<unknown>, line 144)
+WARNING:tensorflow:ModelCheckpoint mode 1 is unknown, fallback to auto mode.
+WARNING:tensorflow:Model failed to serialize as JSON. Ignoring... Layers with arguments in `__init__` must override `get_config`.
+2020-06-15 17:18:23.031977: I tensorflow/core/profiler/lib/profiler_session.cc:225] Profiler session started.
+2020-06-15 17:18:23.032275: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libcupti.so.10.1'; dlerror: libcupti.so.10.1: cannot open shared object file: No such file or directory
+2020-06-15 17:18:23.032314: E tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1307] function cupti_interface_->Subscribe( &subscriber_, (CUpti_CallbackFunc)ApiCallback, this)failed with error CUPTI could not be loaded or symbol could not be found.
+2020-06-15 17:18:23.032330: E tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1346] function cupti_interface_->ActivityRegisterCallbacks( AllocCuptiActivityBuffer, FreeCuptiActivityBuffer)failed with error CUPTI could not be loaded or symbol could not be found.
+2020-06-15 17:18:26.461281: E tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1329] function cupti_interface_->EnableCallback( 0 , subscriber_, CUPTI_CB_DOMAIN_DRIVER_API, cbid)failed with error CUPTI could not be loaded or symbol could not be found.
+2020-06-15 17:18:26.461372: I tensorflow/core/profiler/internal/gpu/device_tracer.cc:88]  GpuTracer has collected 0 callback api events and 0 activity events.
+2020-06-15 17:21:21.898209: W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled
+2020-06-15 17:24:27.449155: W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled
+2020-06-15 17:27:34.489737: W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled
+2020-06-15 17:30:39.394709: W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled
+2020-06-15 17:33:52.841270: W tensorflow/core/kernels/data/generator_dataset_op.cc:103] Error occurred when finalizing GeneratorDataset iterator: Cancelled: Operation was cancelled

Result_yolo3_fault_2/yolo3_full_yolo.output

@@ -0,0 +1,337 @@
+Seen labels: 	{'2': 127}
+
+Given labels: 	['2']
+
+Training on: 	['2']
+
+multi_gpu:1
+
+Loading pretrained weights.
+
+Train for 40 steps, validate for 10 steps
+Loading pretrained weights.
+
+Epoch 1/200
+ - 269s - loss: 10.3918 - yolo_layer_1_loss: 0.2670 - yolo_layer_2_loss: 1.2377 - yolo_layer_3_loss: 8.8871
+
+Epoch 00001: loss improved from inf to 10.39175, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 2/200
+ - 248s - loss: 9.0657 - yolo_layer_1_loss: 0.0042 - yolo_layer_2_loss: 0.8107 - yolo_layer_3_loss: 8.2509
+
+Epoch 00002: loss improved from 10.39175 to 9.06572, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 3/200
+ - 250s - loss: 7.7642 - yolo_layer_1_loss: 0.0021 - yolo_layer_2_loss: 0.5234 - yolo_layer_3_loss: 7.2387
+
+Epoch 00003: loss improved from 9.06572 to 7.76417, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 4/200
+ - 250s - loss: 7.6795 - yolo_layer_1_loss: 0.0018 - yolo_layer_2_loss: 0.2856 - yolo_layer_3_loss: 7.3921
+
+Epoch 00004: loss improved from 7.76417 to 7.67951, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 5/200
+ - 250s - loss: 7.8414 - yolo_layer_1_loss: 0.0011 - yolo_layer_2_loss: 0.4949 - yolo_layer_3_loss: 7.3454
+
+Epoch 00005: loss did not improve from 7.67951
+Epoch 6/200
+ - 250s - loss: 8.1152 - yolo_layer_1_loss: 0.0016 - yolo_layer_2_loss: 0.5053 - yolo_layer_3_loss: 7.6083
+
+Epoch 00006: loss did not improve from 7.67951
+Epoch 7/200
+ - 249s - loss: 7.3296 - yolo_layer_1_loss: 6.0819e-04 - yolo_layer_2_loss: 0.4263 - yolo_layer_3_loss: 6.9027
+
+Epoch 00007: loss improved from 7.67951 to 7.32961, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 8/200
+ - 250s - loss: 7.8274 - yolo_layer_1_loss: 7.9464e-04 - yolo_layer_2_loss: 0.5090 - yolo_layer_3_loss: 7.3175
+
+Epoch 00008: loss did not improve from 7.32961
+Epoch 9/200
+ - 250s - loss: 8.0380 - yolo_layer_1_loss: 8.6137e-04 - yolo_layer_2_loss: 0.6514 - yolo_layer_3_loss: 7.3856
+
+Epoch 00009: loss did not improve from 7.32961
+Epoch 10/200
+ - 250s - loss: 7.9667 - yolo_layer_1_loss: 0.0022 - yolo_layer_2_loss: 0.6072 - yolo_layer_3_loss: 7.3574
+
+Epoch 00010: loss did not improve from 7.32961
+Epoch 11/200
+ - 250s - loss: 8.4253 - yolo_layer_1_loss: 0.0012 - yolo_layer_2_loss: 0.6697 - yolo_layer_3_loss: 7.7544
+
+Epoch 00011: loss did not improve from 7.32961
+Epoch 12/200
+ - 250s - loss: 8.0104 - yolo_layer_1_loss: 0.0010 - yolo_layer_2_loss: 0.2845 - yolo_layer_3_loss: 7.7248
+
+Epoch 00012: loss did not improve from 7.32961
+Epoch 13/200
+ - 250s - loss: 8.7709 - yolo_layer_1_loss: 7.4243e-04 - yolo_layer_2_loss: 0.8608 - yolo_layer_3_loss: 7.9093
+
+Epoch 00013: loss did not improve from 7.32961
+Epoch 14/200
+ - 250s - loss: 8.7679 - yolo_layer_1_loss: 7.1667e-04 - yolo_layer_2_loss: 1.2540 - yolo_layer_3_loss: 7.5131
+
+Epoch 00014: loss did not improve from 7.32961
+Epoch 15/200
+ - 249s - loss: 6.9143 - yolo_layer_1_loss: 6.4281e-04 - yolo_layer_2_loss: 0.8410 - yolo_layer_3_loss: 6.0726
+
+Epoch 00015: loss improved from 7.32961 to 6.91428, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 16/200
+ - 250s - loss: 6.4881 - yolo_layer_1_loss: 5.6383e-04 - yolo_layer_2_loss: 0.4001 - yolo_layer_3_loss: 6.0874
+
+Epoch 00016: loss improved from 6.91428 to 6.48809, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 17/200
+ - 250s - loss: 8.0069 - yolo_layer_1_loss: 4.0944e-04 - yolo_layer_2_loss: 1.3397 - yolo_layer_3_loss: 6.6669
+
+Epoch 00017: loss did not improve from 6.48809
+Epoch 18/200
+ - 250s - loss: 6.3577 - yolo_layer_1_loss: 2.9129e-04 - yolo_layer_2_loss: 0.7906 - yolo_layer_3_loss: 5.5668
+
+Epoch 00018: loss improved from 6.48809 to 6.35768, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 19/200
+ - 249s - loss: 6.7679 - yolo_layer_1_loss: 3.0164e-04 - yolo_layer_2_loss: 0.5212 - yolo_layer_3_loss: 6.2463
+
+Epoch 00019: loss did not improve from 6.35768
+Epoch 20/200
+ - 249s - loss: 6.4508 - yolo_layer_1_loss: 2.7047e-04 - yolo_layer_2_loss: 0.2659 - yolo_layer_3_loss: 6.1846
+
+Epoch 00020: loss did not improve from 6.35768
+Epoch 21/200
+ - 249s - loss: 6.3571 - yolo_layer_1_loss: 2.9117e-04 - yolo_layer_2_loss: 0.3484 - yolo_layer_3_loss: 6.0084
+
+Epoch 00021: loss improved from 6.35768 to 6.35713, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 22/200
+ - 249s - loss: 6.4931 - yolo_layer_1_loss: 2.4105e-04 - yolo_layer_2_loss: 0.5846 - yolo_layer_3_loss: 5.9083
+
+Epoch 00022: loss did not improve from 6.35713
+Epoch 23/200
+ - 249s - loss: 6.4707 - yolo_layer_1_loss: 1.9136e-04 - yolo_layer_2_loss: 0.5451 - yolo_layer_3_loss: 5.9254
+
+Epoch 00023: loss did not improve from 6.35713
+Epoch 24/200
+ - 249s - loss: 5.9562 - yolo_layer_1_loss: 1.3915e-04 - yolo_layer_2_loss: 0.2577 - yolo_layer_3_loss: 5.6984
+
+Epoch 00024: loss improved from 6.35713 to 5.95620, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 25/200
+ - 249s - loss: 6.1448 - yolo_layer_1_loss: 1.2286e-04 - yolo_layer_2_loss: 0.6805 - yolo_layer_3_loss: 5.4642
+
+Epoch 00025: loss did not improve from 5.95620
+Epoch 26/200
+ - 249s - loss: 5.9585 - yolo_layer_1_loss: 1.0858e-04 - yolo_layer_2_loss: 0.4485 - yolo_layer_3_loss: 5.5098
+
+Epoch 00026: loss did not improve from 5.95620
+Epoch 27/200
+ - 249s - loss: 6.1453 - yolo_layer_1_loss: 9.7766e-05 - yolo_layer_2_loss: 0.5231 - yolo_layer_3_loss: 5.6221
+
+Epoch 00027: loss did not improve from 5.95620
+Epoch 28/200
+ - 249s - loss: 5.6772 - yolo_layer_1_loss: 1.9072e-04 - yolo_layer_2_loss: 0.2765 - yolo_layer_3_loss: 5.4006
+
+Epoch 00028: loss improved from 5.95620 to 5.67724, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 29/200
+ - 249s - loss: 5.3301 - yolo_layer_1_loss: 1.0635e-04 - yolo_layer_2_loss: 8.7561e-04 - yolo_layer_3_loss: 5.3291
+
+Epoch 00029: loss improved from 5.67724 to 5.33012, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 30/200
+ - 249s - loss: 6.8329 - yolo_layer_1_loss: 1.4615e-04 - yolo_layer_2_loss: 1.5756 - yolo_layer_3_loss: 5.2572
+
+Epoch 00030: loss did not improve from 5.33012
+Epoch 31/200
+ - 249s - loss: 5.9033 - yolo_layer_1_loss: 1.9254e-04 - yolo_layer_2_loss: 0.8525 - yolo_layer_3_loss: 5.0506
+
+Epoch 00031: loss did not improve from 5.33012
+Epoch 32/200
+ - 249s - loss: 5.2796 - yolo_layer_1_loss: 1.4435e-04 - yolo_layer_2_loss: 0.2685 - yolo_layer_3_loss: 5.0109
+
+Epoch 00032: loss improved from 5.33012 to 5.27958, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 33/200
+ - 249s - loss: 6.2346 - yolo_layer_1_loss: 6.7379e-05 - yolo_layer_2_loss: 1.3654 - yolo_layer_3_loss: 4.8691
+
+Epoch 00033: loss did not improve from 5.27958
+Epoch 34/200
+ - 249s - loss: 6.1787 - yolo_layer_1_loss: 6.0842e-05 - yolo_layer_2_loss: 0.6770 - yolo_layer_3_loss: 5.5016
+
+Epoch 00034: loss did not improve from 5.27958
+Epoch 35/200
+ - 249s - loss: 5.2614 - yolo_layer_1_loss: 6.5419e-05 - yolo_layer_2_loss: 0.1292 - yolo_layer_3_loss: 5.1321
+
+Epoch 00035: loss improved from 5.27958 to 5.26138, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 36/200
+ - 249s - loss: 5.2841 - yolo_layer_1_loss: 2.1123e-04 - yolo_layer_2_loss: 0.1537 - yolo_layer_3_loss: 5.1302
+
+Epoch 00036: loss did not improve from 5.26138
+Epoch 37/200
+ - 250s - loss: 5.5240 - yolo_layer_1_loss: 3.0572e-04 - yolo_layer_2_loss: 0.5670 - yolo_layer_3_loss: 4.9567
+
+Epoch 00037: loss did not improve from 5.26138
+Epoch 38/200
+ - 250s - loss: 5.7255 - yolo_layer_1_loss: 1.5723e-04 - yolo_layer_2_loss: 0.6625 - yolo_layer_3_loss: 5.0628
+
+Epoch 00038: loss did not improve from 5.26138
+Epoch 39/200
+ - 249s - loss: 6.1184 - yolo_layer_1_loss: 1.5941e-04 - yolo_layer_2_loss: 0.7594 - yolo_layer_3_loss: 5.3588
+
+Epoch 00039: loss did not improve from 5.26138
+Epoch 40/200
+ - 250s - loss: 6.0649 - yolo_layer_1_loss: 1.6101e-04 - yolo_layer_2_loss: 0.6664 - yolo_layer_3_loss: 5.3983
+
+Epoch 00040: loss did not improve from 5.26138
+Epoch 41/200
+ - 249s - loss: 7.6413 - yolo_layer_1_loss: 1.6196e-04 - yolo_layer_2_loss: 2.1583 - yolo_layer_3_loss: 5.4829
+
+Epoch 00041: loss did not improve from 5.26138
+Epoch 42/200
+ - 249s - loss: 6.7180 - yolo_layer_1_loss: 1.1377e-04 - yolo_layer_2_loss: 1.2027 - yolo_layer_3_loss: 5.5152
+
+Epoch 00042: loss did not improve from 5.26138
+Epoch 43/200
+ - 250s - loss: 5.6705 - yolo_layer_1_loss: 9.3619e-05 - yolo_layer_2_loss: 0.7231 - yolo_layer_3_loss: 4.9473
+
+Epoch 00043: loss did not improve from 5.26138
+Epoch 44/200
+ - 249s - loss: 6.2414 - yolo_layer_1_loss: 1.0592e-04 - yolo_layer_2_loss: 0.6340 - yolo_layer_3_loss: 5.6073
+
+Epoch 00044: loss did not improve from 5.26138
+Epoch 45/200
+ - 249s - loss: 5.8108 - yolo_layer_1_loss: 5.7582e-05 - yolo_layer_2_loss: 0.6373 - yolo_layer_3_loss: 5.1734
+
+Epoch 00045: loss did not improve from 5.26138
+Epoch 46/200
+ - 249s - loss: 6.9616 - yolo_layer_1_loss: 9.3750e-05 - yolo_layer_2_loss: 0.8949 - yolo_layer_3_loss: 6.0666
+
+Epoch 00046: loss did not improve from 5.26138
+Epoch 47/200
+ - 249s - loss: 5.3763 - yolo_layer_1_loss: 8.0919e-05 - yolo_layer_2_loss: 0.2562 - yolo_layer_3_loss: 5.1200
+
+Epoch 00047: loss did not improve from 5.26138
+Epoch 48/200
+ - 249s - loss: 5.4027 - yolo_layer_1_loss: 5.5944e-05 - yolo_layer_2_loss: 0.6505 - yolo_layer_3_loss: 4.7521
+
+Epoch 00048: loss did not improve from 5.26138
+Epoch 49/200
+ - 249s - loss: 5.4222 - yolo_layer_1_loss: 5.0298e-05 - yolo_layer_2_loss: 0.5950 - yolo_layer_3_loss: 4.8272
+
+Epoch 00049: loss did not improve from 5.26138
+Epoch 50/200
+ - 249s - loss: 5.5717 - yolo_layer_1_loss: 5.9565e-05 - yolo_layer_2_loss: 0.1377 - yolo_layer_3_loss: 5.4340
+
+Epoch 00050: loss did not improve from 5.26138
+
+Epoch 00050: ReduceLROnPlateau reducing learning rate to 0.0005000000237487257.
+Epoch 51/200
+ - 249s - loss: 5.9566 - yolo_layer_1_loss: 6.9738e-05 - yolo_layer_2_loss: 0.4066 - yolo_layer_3_loss: 5.5499
+
+Epoch 00051: loss did not improve from 5.26138
+Epoch 52/200
+ - 249s - loss: 4.5897 - yolo_layer_1_loss: 5.2179e-05 - yolo_layer_2_loss: 0.4062 - yolo_layer_3_loss: 4.1835
+
+Epoch 00052: loss improved from 5.26138 to 4.58969, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 53/200
+ - 249s - loss: 4.8618 - yolo_layer_1_loss: 5.1721e-05 - yolo_layer_2_loss: 0.5580 - yolo_layer_3_loss: 4.3037
+
+Epoch 00053: loss did not improve from 4.58969
+Epoch 54/200
+ - 249s - loss: 4.8267 - yolo_layer_1_loss: 5.7835e-05 - yolo_layer_2_loss: 0.3104 - yolo_layer_3_loss: 4.5163
+
+Epoch 00054: loss did not improve from 4.58969
+Epoch 55/200
+ - 249s - loss: 5.7597 - yolo_layer_1_loss: 7.0499e-05 - yolo_layer_2_loss: 0.8402 - yolo_layer_3_loss: 4.9194
+
+Epoch 00055: loss did not improve from 4.58969
+Epoch 56/200
+ - 249s - loss: 4.5171 - yolo_layer_1_loss: 4.7918e-05 - yolo_layer_2_loss: 0.6464 - yolo_layer_3_loss: 3.8707
+
+Epoch 00056: loss improved from 4.58969 to 4.51712, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 57/200
+ - 249s - loss: 4.8879 - yolo_layer_1_loss: 4.4887e-05 - yolo_layer_2_loss: 0.3918 - yolo_layer_3_loss: 4.4960
+
+Epoch 00057: loss did not improve from 4.51712
+Epoch 58/200
+ - 250s - loss: 4.6380 - yolo_layer_1_loss: 4.3016e-05 - yolo_layer_2_loss: 0.4356 - yolo_layer_3_loss: 4.2024
+
+Epoch 00058: loss did not improve from 4.51712
+Epoch 59/200
+ - 249s - loss: 5.1633 - yolo_layer_1_loss: 3.9667e-05 - yolo_layer_2_loss: 0.7258 - yolo_layer_3_loss: 4.4374
+
+Epoch 00059: loss did not improve from 4.51712
+Epoch 60/200
+ - 249s - loss: 4.8450 - yolo_layer_1_loss: 4.7864e-05 - yolo_layer_2_loss: 0.8874 - yolo_layer_3_loss: 3.9576
+
+Epoch 00060: loss did not improve from 4.51712
+Epoch 61/200
+ - 249s - loss: 4.7428 - yolo_layer_1_loss: 4.3229e-05 - yolo_layer_2_loss: 0.2623 - yolo_layer_3_loss: 4.4805
+
+Epoch 00061: loss did not improve from 4.51712
+Epoch 62/200
+ - 249s - loss: 5.4766 - yolo_layer_1_loss: 4.1361e-05 - yolo_layer_2_loss: 0.9122 - yolo_layer_3_loss: 4.5644
+
+Epoch 00062: loss did not improve from 4.51712
+Epoch 63/200
+ - 249s - loss: 4.2940 - yolo_layer_1_loss: 4.3164e-05 - yolo_layer_2_loss: 0.5223 - yolo_layer_3_loss: 3.7716
+
+Epoch 00063: loss improved from 4.51712 to 4.29402, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 64/200
+ - 249s - loss: 4.6400 - yolo_layer_1_loss: 5.8545e-05 - yolo_layer_2_loss: 0.6264 - yolo_layer_3_loss: 4.0135
+
+Epoch 00064: loss did not improve from 4.29402
+Epoch 65/200
+ - 249s - loss: 4.5991 - yolo_layer_1_loss: 5.0934e-05 - yolo_layer_2_loss: 0.4342 - yolo_layer_3_loss: 4.1649
+
+Epoch 00065: loss did not improve from 4.29402
+Epoch 66/200
+ - 249s - loss: 4.9235 - yolo_layer_1_loss: 5.6441e-05 - yolo_layer_2_loss: 0.9522 - yolo_layer_3_loss: 3.9712
+
+Epoch 00066: loss did not improve from 4.29402
+Epoch 67/200
+ - 249s - loss: 5.0479 - yolo_layer_1_loss: 5.0447e-05 - yolo_layer_2_loss: 0.9907 - yolo_layer_3_loss: 4.0571
+
+Epoch 00067: loss did not improve from 4.29402
+Epoch 68/200
+ - 250s - loss: 4.8168 - yolo_layer_1_loss: 4.6199e-05 - yolo_layer_2_loss: 1.1044 - yolo_layer_3_loss: 3.7123
+
+Epoch 00068: loss did not improve from 4.29402
+Epoch 69/200
+ - 249s - loss: 4.5846 - yolo_layer_1_loss: 4.9725e-05 - yolo_layer_2_loss: 0.2604 - yolo_layer_3_loss: 4.3242
+
+Epoch 00069: loss did not improve from 4.29402
+Epoch 70/200
+ - 249s - loss: 5.3970 - yolo_layer_1_loss: 3.2394e-05 - yolo_layer_2_loss: 0.7830 - yolo_layer_3_loss: 4.6140
+
+Epoch 00070: loss did not improve from 4.29402
+Epoch 71/200
+ - 249s - loss: 4.5713 - yolo_layer_1_loss: 3.9810e-05 - yolo_layer_2_loss: 0.4039 - yolo_layer_3_loss: 4.1674
+
+Epoch 00071: loss did not improve from 4.29402
+Epoch 72/200
+ - 249s - loss: 3.8602 - yolo_layer_1_loss: 3.7398e-05 - yolo_layer_2_loss: 0.7578 - yolo_layer_3_loss: 3.1023
+
+Epoch 00072: loss improved from 4.29402 to 3.86019, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+Epoch 73/200
+ - 249s - loss: 4.2252 - yolo_layer_1_loss: 3.9543e-05 - yolo_layer_2_loss: 0.1264 - yolo_layer_3_loss: 4.0988
+
+Epoch 00073: loss did not improve from 3.86019
+Epoch 74/200
+ - 249s - loss: 4.4763 - yolo_layer_1_loss: 3.6080e-05 - yolo_layer_2_loss: 0.5650 - yolo_layer_3_loss: 3.9113
+
+Epoch 00074: loss did not improve from 3.86019
+Epoch 75/200
+ - 249s - loss: 3.5016 - yolo_layer_1_loss: 3.7154e-05 - yolo_layer_2_loss: 0.6020 - yolo_layer_3_loss: 2.8996
+
+Epoch 00075: loss improved from 3.86019 to 3.50156, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+
+Epoch 00076: loss improved from 3.50156 to 4.22446, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+40/40 - 183s - loss: 4.2245 - yolo_layer_loss: 6.4795e-05 - yolo_layer_1_loss: 0.6512 - yolo_layer_2_loss: 3.5732 - val_loss: 6.4259 - val_yolo_layer_loss: 6.5558e-05 - val_yolo_layer_1_loss: 1.3530 - val_yolo_layer_2_loss: 5.0729
+Epoch 2/5
+
+Epoch 00077: loss did not improve from 4.22446
+40/40 - 186s - loss: 4.8331 - yolo_layer_loss: 4.1540e-05 - yolo_layer_1_loss: 1.0007 - yolo_layer_2_loss: 3.8323 - val_loss: 4.9868 - val_yolo_layer_loss: 2.5803e-05 - val_yolo_layer_1_loss: 1.2658e-04 - val_yolo_layer_2_loss: 4.9867
+Epoch 3/5
+
+Epoch 00078: loss did not improve from 4.22446
+40/40 - 187s - loss: 4.3298 - yolo_layer_loss: 8.7633e-06 - yolo_layer_1_loss: 0.6481 - yolo_layer_2_loss: 3.6816 - val_loss: 6.7895 - val_yolo_layer_loss: 8.8222e-06 - val_yolo_layer_1_loss: 0.5092 - val_yolo_layer_2_loss: 6.2803
+Epoch 4/5
+
+Epoch 00079: loss did not improve from 4.22446
+40/40 - 185s - loss: 4.6490 - yolo_layer_loss: 6.7689e-06 - yolo_layer_1_loss: 0.8990 - yolo_layer_2_loss: 3.7501 - val_loss: 5.3483 - val_yolo_layer_loss: 6.3485e-06 - val_yolo_layer_1_loss: 9.0986e-05 - val_yolo_layer_2_loss: 5.3482
+Epoch 5/5
+
+Epoch 00080: loss improved from 4.22446 to 3.98889, saving model to Result_yolo3_fault_2/yolo3_full_fault_2.h5
+40/40 - 193s - loss: 3.9889 - yolo_layer_loss: 5.2988e-06 - yolo_layer_1_loss: 0.6052 - yolo_layer_2_loss: 3.3837 - val_loss: 6.2318 - val_yolo_layer_loss: 4.7080e-06 - val_yolo_layer_1_loss: 6.1751e-05 - val_yolo_layer_2_loss: 6.2318
+35 instances of class 2 with average precision: 0.7193
+mAP using the weighted average of precisions among classes: 0.7193
+mAP: 0.7193

Result_yolo3_fault_2/yolo3_full_yolo_test.err

@@ -0,0 +1,11 @@
+2020-06-15 17:34:31.100582: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer.so.6'; dlerror: libnvinfer.so.6: cannot open shared object file: No such file or directory
+2020-06-15 17:34:31.100864: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libnvinfer_plugin.so.6'; dlerror: libnvinfer_plugin.so.6: cannot open shared object file: No such file or directory
+2020-06-15 17:34:31.100885: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:30] Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
+2020-06-15 17:34:32.742430: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory
+2020-06-15 17:34:32.742474: E tensorflow/stream_executor/cuda/cuda_driver.cc:351] failed call to cuInit: UNKNOWN ERROR (303)
+2020-06-15 17:34:32.742498: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:156] kernel driver does not appear to be running on this host (dlsaavedra-X406UAR): /proc/driver/nvidia/version does not exist
+2020-06-15 17:34:32.794751: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
+2020-06-15 17:34:32.823902: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 1800000000 Hz
+2020-06-15 17:34:32.825891: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x560fce206ff0 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
+2020-06-15 17:34:32.825979: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
+WARNING:tensorflow:No training configuration found in save file: the model was *not* compiled. Compile it manually.

Result_yolo3_fault_2/yolo3_full_yolo_test.output

@@ -0,0 +1,4 @@
+dict_items([(0, (0.7269822428245609, 35.0))])
+35 instances of class 2 with average precision: 0.7270
+mAP using the weighted average of precisions among classes: 0.7270
+mAP: 0.7270

config_full_yolo_fault_2_infer.json

@@ -3,7 +3,7 @@
         "min_input_size":       400,
         "max_input_size":       400,
         "anchors":              [5,7, 10,14, 15, 15, 26,32, 45,119, 54,18, 94,59, 109,183, 200,21],
-        "labels":               ["1"],
+        "labels":               ["2"],
 	"backend": 		"keras-yolo3-master/full_yolo_backend.h5"
     },
 
@@ -12,22 +12,22 @@
         "train_annot_folder":   "Train&Test_H/Train/anns/",
 	"cache_name":           "Result_yolo3_fault_2/experimento_fault_2_gpu.pkl",
 
-        "train_times":          1,
+  "train_times":          1,
 
-        "batch_size":           2,
-        "learning_rate":        1e-4,
-        "nb_epochs":            200,
-        "warmup_epochs":        15,
-        "ignore_thresh":        0.5,
-        "gpus":                 "0,1",
+  "batch_size":           2,
+  "learning_rate":        1e-4,
+  "nb_epochs":            200,
+  "warmup_epochs":        10,
+  "ignore_thresh":        0.5,
+  "gpus":                 "0",
 
-	"grid_scales":          [1,1,1],
-        "obj_scale":            5,
-        "noobj_scale":          1,
-        "xywh_scale":           1,
-        "class_scale":          1,
+"grid_scales":          [1,1,1],
+  "obj_scale":            5,
+  "noobj_scale":          1,
+  "xywh_scale":           1,
+  "class_scale":          1,
 
-	"tensorboard_dir":      "log_experimento_fault_gpu_2",
+  "tensorboard_dir":      "Result_yolo3_fault_2/log_experimento_fault_gpu_2",
 	"saved_weights_name":   "Result_yolo3_fault_2/yolo3_full_fault_2.h5",
         "debug":                true
     },

config_full_yolo_fault_2_train.json

@@ -3,7 +3,7 @@
         "min_input_size":       400,
         "max_input_size":       400,
         "anchors":              [5,7, 10,14, 15, 15, 26,32, 45,119, 54,18, 94,59, 109,183, 200,21],
-        "labels":               ["1"],
+        "labels":               ["2"],
 	"backend": 		"keras-yolo3-master/full_yolo_backend.h5"
     },
 
@@ -17,9 +17,9 @@
         "batch_size":           2,
         "learning_rate":        1e-4,
         "nb_epochs":            200,
-        "warmup_epochs":        15,
+        "warmup_epochs":        10,
         "ignore_thresh":        0.5,
-        "gpus":                 "0,1",
+        "gpus":                 "0",
 
 	"grid_scales":          [1,1,1],
         "obj_scale":            5,
@@ -27,7 +27,7 @@
         "xywh_scale":           1,
         "class_scale":          1,
 
-	"tensorboard_dir":      "log_experimento_fault_gpu_2",
+	"tensorboard_dir":      "Result_yolo3_fault_2/log_experimento_fault_gpu_2",
 	"saved_weights_name":   "Result_yolo3_fault_2/yolo3_full_fault_2.h5",
         "debug":                true
     },

keras-yolo3-master/train.py

@@ -69,7 +69,7 @@ def create_callbacks(saved_weights_name, tensorboard_logs, model_to_save):
     makedirs(tensorboard_logs)
 
     early_stop = EarlyStopping(
-        monitor     = 'val_loss',
+        monitor     = 'loss',
         min_delta   = 0.01,
         patience    = 25,
         mode        = 'min',
@@ -85,13 +85,14 @@ def create_callbacks(saved_weights_name, tensorboard_logs, model_to_save):
         save_freq       = 1
     )"""
     checkpoint = ModelCheckpoint(filepath=saved_weights_name,
-                                 monitor='val_loss',
+                                 monitor='loss',
                                  save_best_only=True,
                                  save_weights_only=True,
+                                 mode = 1,
                                  verbose=1)
 
     reduce_on_plateau = ReduceLROnPlateau(
-        monitor  = 'val_loss',
+        monitor  = 'loss',
         factor   = 0.5,
         patience = 15,
         verbose  = 1,
@@ -105,7 +106,7 @@ def create_callbacks(saved_weights_name, tensorboard_logs, model_to_save):
         write_graph            = True,
         write_images           = True,
     )
-    return [early_stop, checkpoint, reduce_on_plateau]
+    return [early_stop, checkpoint, reduce_on_plateau, tensorboard]
 
 def create_model(
     nb_class,
@@ -273,9 +274,10 @@ def _main_(args):
 
 
     # make a GPU version of infer_model for evaluation
-    if multi_gpu > 1:
-        infer_model = load_model(config['train']['saved_weights_name'])
-
+    #if multi_gpu > 1:
+    #    infer_model = load_model(config['train']['saved_weights_name'])
+    infer_model.load_weights(config['train']['saved_weights_name'])
+    infer_model.save(config['train']['saved_weights_name'])
     ###############################
     #   Run the evaluation
     ###############################