python /home/admin/mtr/script_for_cron.py -j datou_current3 -m 10 -a ' -a 4234 -l 5 ' -s datou_current_4234 -M 0 -S 0 -U 95,95,80 import MySQLdb succeeded Import error (python version) ['/Users/moilerat/Documents/Fotonower/install/caffe/distribute/python', '/home/admin/workarea/git/Velours/python/prod', '/home/admin/workarea/install/caffe_cuda8_python3/python', '/home/admin/workarea/install/darknet', '/home/admin/workarea/git/Velours/python', '/home/admin/workarea/install/caffe_frcnn_python3/py-faster-rcnn/caffe-fast-rcnn/python', '/home/admin/mtr/.credentials', '/home/admin/workarea/install/caffe/python', '/home/admin/workarea/install/caffe_frcnn/py-faster-rcnn/tools', '/home/admin/workarea/git/fotonowerpip', '/home/admin/workarea/install/segment-anything', '/home/admin/workarea/git/pyfvs', '/usr/lib/python38.zip', '/usr/lib/python3.8', '/usr/lib/python3.8/lib-dynload', '/home/admin/.local/lib/python3.8/site-packages', '/usr/local/lib/python3.8/dist-packages', '/usr/lib/python3/dist-packages'] process id : 1166765 load datou : 4234 # VR 17-11-17 : to create in DB ! Here we check the datou graph and we reorder steps ! Tree builded and cycle checked, now we need to re-order the steps ! We have currenlty an error because there is no dependence between the last step for the case tile - detect - glue We can either keep the depence of, it is better to keep an order compatible with the id of steps if we do not have sons, so a lexical order : (number_son, step_id) All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! DONE and to test : checkNoCycle ! Here we check the consistency of inputs/outputs number between the given ones and the db ! eke 1-6-18 : checkConsistencyNbInputNbOutput should be processed after step reordering ! WARNING : number of outputs for step 11415 rle_unique_nms_with_priority is not consistent : 2 used against 1 in the step definition ! Step 11419 crop_condition have less inputs used (1) than in the step definition (2) : maybe we manage optionnal inputs ! WARNING : number of outputs for step 11419 crop_condition is not consistent : 4 used against 3 in the step definition ! WARNING : number of inputs for step 11416 ventilate_hashtags_in_portfolio is not consistent : 2 used against 1 in the step definition ! WARNING : number of outputs for step 11416 ventilate_hashtags_in_portfolio is not consistent : 2 used against 1 in the step definition ! Step 11417 final have less inputs used (2) than in the step definition (3) : maybe we manage optionnal inputs ! Step 11417 final have less outputs used (1) than in the step definition (2) : some outputs may be not used ! Step 11422 split_time_score have less inputs used (1) than in the step definition (2) : maybe we manage optionnal inputs ! Number of inputs / outputs for each step checked ! Here we check the consistency of outputs/inputs types during steps connections eke 1-6-18 : checkConsistencyTypeOutputInput should be processed after checkConsistencyNbInputNbOutput ! We ignore checkConsistencyTypeOutputInput for datou_step final ! WARNING : output 0 of step 11416 have datatype=10 whereas input 2 of step 11418 have datatype=6 We ignore checkConsistencyTypeOutputInput for datou_step final ! WARNING : type of input 2 of step 11419 doesn't seem to be define in the database( WARNING : output 1 of step 11415 have datatype=7 whereas input 2 of step 11419 have datatype=None WARNING : type of output 3 of step 11419 doesn't seem to be define in the database( WARNING : type of input 1 of step 11416 doesn't seem to be define in the database( WARNING : type of output 1 of step 11416 doesn't seem to be define in the database( WARNING : type of input 3 of step 11417 doesn't seem to be define in the database( We ignore checkConsistencyTypeOutputInput for datou_step final ! WARNING : output 0 of step 11416 have datatype=10 whereas input 0 of step 11420 have datatype=18 WARNING : type of input 5 of step 11418 doesn't seem to be define in the database( WARNING : output 0 of step 11420 have datatype=11 whereas input 5 of step 11418 have datatype=None WARNING : type of input 2 of step 11416 doesn't seem to be define in the database( WARNING : output 0 of step 11421 have datatype=5 whereas input 2 of step 11416 have datatype=None WARNING : output 0 of step 11418 have datatype=10 whereas input 0 of step 11422 have datatype=18 DataTypes for each output/input checked ! Unexpected type for variable list_input_json ERROR or WARNING : can't parse json string Expecting value: line 1 column 1 (char 0) Tried to parse : chemin de la photo was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? [(photo_id, hashtag_id, hashtag_type, x0, x1, y0, y1, score, seg_temp, polygons), ...] was removed should we ? chemin de la photo was removed should we ? chemin de la photo was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? chemin de la photo was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? chemin de la photo was removed should we ? donnée sous forme de texte was removed should we ? [ (photo_id, photo_id_loc, hashtag_type, x0, x1, y0, y1, score), ...] was removed should we ? [ (photo_id_loc, hashtag_id, hashtag_type, x0, x1, y0, y1, score, None), ...] was removed should we ? chemin de la photo was removed should we ? id de la photo (peut être local ou global) was removed should we ? chemin de la photo was removed should we ? (x0, y0, x1, y1) was removed should we ? chemin de la photo was removed should we ? [ (photo_id, hashtag_id_0, score_0), (photo_id, hashtag_id_1, score_1), ...] was removed should we ? output optionnel thcl was removed should we ? None was removed should we ? donnée sous forme de texte was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? id de la photo (peut être local ou global) was removed should we ? donnée sous forme de texte was removed should we ? donnée sous forme de texte was removed should we ? donnée sous forme de texte was removed should we ? None was removed should we ? donnée sous forme de nombre was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? (photo_id, hashtag_id, score_max) was removed should we ? donnée sous forme de texte was removed should we ? None was removed should we ? donnée sous forme de texte was removed should we ? [ptf_id0,ptf_id1...] was removed should we ? FOUND : 1 Here is data_from_sql_as_vec to set the ParamDescriptorType : (5275, 'learn_RUBBIA_REFUS_AMIENS_23', 16384, 25088, 'learn_RUBBIA_REFUS_AMIENS_23', 'pool5', 10.0, None, None, 256, None, 0, None, 8, None, None, -1000.0, 1, datetime.datetime(2021, 4, 23, 14, 19, 39), datetime.datetime(2021, 4, 23, 14, 19, 39)) FOUND : 1 Here is data_from_sql_as_vec to set the ParamDescriptorType : (5561, 'learn_rubbia_refus_2500', 2048, 2048, 'learn_rubbia_refus_2500', 'pool5', 10.0, None, None, 256, None, 0, None, 8, None, None, -1000.0, 3, datetime.datetime(2021, 12, 2, 19, 10, 8), datetime.datetime(2021, 12, 2, 19, 10, 8)) load thcls load THCL from format json or kwargs add thcl : 2847 in CacheModelConfig load THCL from format json or kwargs add thcl : 3237 in CacheModelConfig load pdts add pdt : 5275 in CacheModelConfig add pdt : 5561 in CacheModelConfig Running datou job : batch_current TODO datou_current to load to do maybe to take outside batchDatouExec updating current state to 1 we have a portfolio with more photos than limit : 35>5 please execute split_portfolio.py -i 21286244 -l 5 size over we load limit photo not treated list_input_json: [] Current got : datou_id : 4234, datou_cur_ids : ['2658182'] with mtr_portfolio_ids : ['21286244'] and first list_photo_ids : [] new path : /proc/1166765/ Inside batchDatouExec : verbose : 0 # VR 17-11-17 : to create in DB ! Here we check the datou graph and we reorder steps ! Tree builded and cycle checked, now we need to re-order the steps ! We have currenlty an error because there is no dependence between the last step for the case tile - detect - glue We can either keep the depence of, it is better to keep an order compatible with the id of steps if we do not have sons, so a lexical order : (number_son, step_id) All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! All sons are already in current list ! DONE and to test : checkNoCycle ! Here we check the consistency of inputs/outputs number between the given ones and the db ! eke 1-6-18 : checkConsistencyNbInputNbOutput should be processed after step reordering ! WARNING : number of outputs for step 11415 rle_unique_nms_with_priority is not consistent : 2 used against 1 in the step definition ! Step 11419 crop_condition have less inputs used (1) than in the step definition (2) : maybe we manage optionnal inputs ! WARNING : number of outputs for step 11419 crop_condition is not consistent : 4 used against 3 in the step definition ! WARNING : number of inputs for step 11416 ventilate_hashtags_in_portfolio is not consistent : 2 used against 1 in the step definition ! WARNING : number of outputs for step 11416 ventilate_hashtags_in_portfolio is not consistent : 2 used against 1 in the step definition ! Step 11417 final have less inputs used (2) than in the step definition (3) : maybe we manage optionnal inputs ! Step 11417 final have less outputs used (1) than in the step definition (2) : some outputs may be not used ! Step 11422 split_time_score have less inputs used (1) than in the step definition (2) : maybe we manage optionnal inputs ! Number of inputs / outputs for each step checked ! Here we check the consistency of outputs/inputs types during steps connections eke 1-6-18 : checkConsistencyTypeOutputInput should be processed after checkConsistencyNbInputNbOutput ! We ignore checkConsistencyTypeOutputInput for datou_step final ! WARNING : output 0 of step 11416 have datatype=10 whereas input 2 of step 11418 have datatype=6 We ignore checkConsistencyTypeOutputInput for datou_step final ! WARNING : type of input 2 of step 11419 doesn't seem to be define in the database( WARNING : output 1 of step 11415 have datatype=7 whereas input 2 of step 11419 have datatype=None WARNING : type of output 3 of step 11419 doesn't seem to be define in the database( WARNING : type of input 1 of step 11416 doesn't seem to be define in the database( WARNING : type of output 1 of step 11416 doesn't seem to be define in the database( WARNING : type of input 3 of step 11417 doesn't seem to be define in the database( We ignore checkConsistencyTypeOutputInput for datou_step final ! WARNING : output 0 of step 11416 have datatype=10 whereas input 0 of step 11420 have datatype=18 WARNING : type of input 5 of step 11418 doesn't seem to be define in the database( WARNING : output 0 of step 11420 have datatype=11 whereas input 5 of step 11418 have datatype=None WARNING : type of input 2 of step 11416 doesn't seem to be define in the database( WARNING : output 0 of step 11421 have datatype=5 whereas input 2 of step 11416 have datatype=None WARNING : output 0 of step 11418 have datatype=10 whereas input 0 of step 11422 have datatype=18 DataTypes for each output/input checked ! List Step Type Loaded in datou : mask_detect, brightness, blur_detection, rle_unique_nms_with_priority, crop_condition, thcl, ventilate_hashtags_in_portfolio, final, velours_tree, send_mail_cod, split_time_score list_input_json : [] origin We have 1 , WARNING: data may be incomplete, need to offset and complete ! BFBFBFBFBFwe have missing 0 photos in the step downloads : photo missing : [] try to delete the photos missing in DB length of list_filenames : 5 ; length of list_pids : 5 ; length of list_args : 5 time to download the photos : 0.6858792304992676 About to test input to load we should then remove the video here, and this would fix the bug of datou_current ! Calling datou_exec Inside datou_exec : verbose : 0 number of steps : 11 step1:mask_detect Tue Mar 11 14:20:30 2025 VR 17-11-17 : now, only for linear exec dependencies tree, some output goes to fill the input of the next VR 22-3-18 : now we test the dependencies tree, but keep two separate code for datou_prepare_output_input until the code is correctly tested, clean and works in both case VR 22-3-18 : but we use the first code for the first step id = -1, build in the code of datou_exec VR 22-3-18 : we should manage here the case when we are at the first step instead of building this step before datou_exec Beginning of datou step mask_detect ! save_polygon : True begin detect begin to check gpu status inside check gpu memory l 3637 free memory gpu now : 10151 max_wait_temp : 1 max_wait : 0 gpu_flag : 0 2025-03-11 14:20:32.972075: I tensorflow/core/platform/cpu_feature_guard.cc:143] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA 2025-03-11 14:20:32.999333: I tensorflow/core/platform/profile_utils/cpu_utils.cc:102] CPU Frequency: 3493065000 Hz 2025-03-11 14:20:33.001124: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x7f5b20000b60 initialized for platform Host (this does not guarantee that XLA will be used). Devices: 2025-03-11 14:20:33.001193: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): Host, Default Version 2025-03-11 14:20:33.004352: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcuda.so.1 2025-03-11 14:20:33.274848: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x10a79150 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices: 2025-03-11 14:20:33.274920: I tensorflow/compiler/xla/service/service.cc:176] StreamExecutor device (0): NVIDIA GeForce RTX 2080 Ti, Compute Capability 7.5 2025-03-11 14:20:33.275808: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties: pciBusID: 0000:41:00.0 name: NVIDIA GeForce RTX 2080 Ti computeCapability: 7.5 coreClock: 1.545GHz coreCount: 68 deviceMemorySize: 10.76GiB deviceMemoryBandwidth: 573.69GiB/s 2025-03-11 14:20:33.276217: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1 2025-03-11 14:20:33.278550: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10 2025-03-11 14:20:33.281154: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10 2025-03-11 14:20:33.281780: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10 2025-03-11 14:20:33.285168: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10 2025-03-11 14:20:33.286386: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10 2025-03-11 14:20:33.291769: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7 2025-03-11 14:20:33.293375: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1703] Adding visible gpu devices: 0 2025-03-11 14:20:33.293505: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1 2025-03-11 14:20:33.294372: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] Device interconnect StreamExecutor with strength 1 edge matrix: 2025-03-11 14:20:33.294389: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1108] 0 2025-03-11 14:20:33.294398: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1121] 0: N 2025-03-11 14:20:33.295994: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1247] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9399 MB memory) -> physical GPU (device: 0, name: NVIDIA GeForce RTX 2080 Ti, pci bus id: 0000:41:00.0, compute capability: 7.5) WARNING:tensorflow:From /home/admin/workarea/git/Velours/python/mtr/mask_rcnn/mask_detection.py:69: The name tf.keras.backend.set_session is deprecated. Please use tf.compat.v1.keras.backend.set_session instead. 2025-03-11 14:20:33.585948: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties: pciBusID: 0000:41:00.0 name: NVIDIA GeForce RTX 2080 Ti computeCapability: 7.5 coreClock: 1.545GHz coreCount: 68 deviceMemorySize: 10.76GiB deviceMemoryBandwidth: 573.69GiB/s 2025-03-11 14:20:33.586189: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1 2025-03-11 14:20:33.586235: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10 2025-03-11 14:20:33.586269: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10 2025-03-11 14:20:33.586298: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10 2025-03-11 14:20:33.586330: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10 2025-03-11 14:20:33.586360: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10 2025-03-11 14:20:33.586391: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7 2025-03-11 14:20:33.588046: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1703] Adding visible gpu devices: 0 2025-03-11 14:20:33.590179: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1561] Found device 0 with properties: pciBusID: 0000:41:00.0 name: NVIDIA GeForce RTX 2080 Ti computeCapability: 7.5 coreClock: 1.545GHz coreCount: 68 deviceMemorySize: 10.76GiB deviceMemoryBandwidth: 573.69GiB/s 2025-03-11 14:20:33.590271: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudart.so.10.1 2025-03-11 14:20:33.590291: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10 2025-03-11 14:20:33.590309: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcufft.so.10 2025-03-11 14:20:33.590326: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcurand.so.10 2025-03-11 14:20:33.590343: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusolver.so.10 2025-03-11 14:20:33.590359: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcusparse.so.10 2025-03-11 14:20:33.590377: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7 2025-03-11 14:20:33.591918: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1703] Adding visible gpu devices: 0 2025-03-11 14:20:33.591976: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1102] Device interconnect StreamExecutor with strength 1 edge matrix: 2025-03-11 14:20:33.591987: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1108] 0 2025-03-11 14:20:33.591995: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1121] 0: N 2025-03-11 14:20:33.593388: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1247] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9399 MB memory) -> physical GPU (device: 0, name: NVIDIA GeForce RTX 2080 Ti, pci bus id: 0000:41:00.0, compute capability: 7.5) Using TensorFlow backend. WARNING:tensorflow:From /home/admin/workarea/install/Mask_RCNN/model.py:396: calling crop_and_resize_v1 (from tensorflow.python.ops.image_ops_impl) with box_ind is deprecated and will be removed in a future version. Instructions for updating: box_ind is deprecated, use box_indices instead WARNING:tensorflow:From /home/admin/workarea/install/Mask_RCNN/model.py:703: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version. Instructions for updating: Use `tf.cast` instead. WARNING:tensorflow:From /home/admin/workarea/install/Mask_RCNN/model.py:729: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version. Instructions for updating: Use `tf.cast` instead. Inside mask_sub_process Inside mask_detect About to load cache.load_thcl_param To do loadFromThcl(), then load ParamDescType : thcl2847 thcls : [{'id': 2847, 'mtr_user_id': 31, 'name': 'learn_RUBBIA_REFUS_AMIENS_23', 'pb_hashtag_id': 0, 'live': b'\x00', 'list_hashtags': 'background,papier,carton,metal,pet_clair,autre,pehd,pet_fonce,environnement', 'svm_portfolios_learning': '0,0,0,0,0,0,0,0,0', 'photo_hashtag_type': 3594, 'photo_desc_type': 5275, 'type_classification': 'mask_rcnn', 'hashtag_id_list': '0,0,0,0,0,0,0,0,0'}] thcl {'id': 2847, 'mtr_user_id': 31, 'name': 'learn_RUBBIA_REFUS_AMIENS_23', 'pb_hashtag_id': 0, 'live': b'\x00', 'list_hashtags': 'background,papier,carton,metal,pet_clair,autre,pehd,pet_fonce,environnement', 'svm_portfolios_learning': '0,0,0,0,0,0,0,0,0', 'photo_hashtag_type': 3594, 'photo_desc_type': 5275, 'type_classification': 'mask_rcnn', 'hashtag_id_list': '0,0,0,0,0,0,0,0,0'} Update svm_hashtag_type_desc : 5275 FOUND : 1 Here is data_from_sql_as_vec to set the ParamDescriptorType : (5275, 'learn_RUBBIA_REFUS_AMIENS_23', 16384, 25088, 'learn_RUBBIA_REFUS_AMIENS_23', 'pool5', 10.0, None, None, 256, None, 0, None, 8, None, None, -1000.0, 1, datetime.datetime(2021, 4, 23, 14, 19, 39), datetime.datetime(2021, 4, 23, 14, 19, 39)) {'thcl': {'id': 2847, 'mtr_user_id': 31, 'name': 'learn_RUBBIA_REFUS_AMIENS_23', 'pb_hashtag_id': 0, 'live': b'\x00', 'list_hashtags': 'background,papier,carton,metal,pet_clair,autre,pehd,pet_fonce,environnement', 'svm_portfolios_learning': '0,0,0,0,0,0,0,0,0', 'photo_hashtag_type': 3594, 'photo_desc_type': 5275, 'type_classification': 'mask_rcnn', 'hashtag_id_list': '0,0,0,0,0,0,0,0,0'}, 'list_hashtags': ['background', 'papier', 'carton', 'metal', 'pet_clair', 'autre', 'pehd', 'pet_fonce', 'environnement'], 'list_hashtags_csv': 'background,papier,carton,metal,pet_clair,autre,pehd,pet_fonce,environnement', 'svm_portfolios_learning': '0,0,0,0,0,0,0,0,0', 'photo_hashtag_type': 3594, 'svm_hashtag_type_desc': 5275, 'photo_desc_type': 5275, 'pb_hashtag_id_or_classifier': 0} list_class_names : ['background', 'papier', 'carton', 'metal', 'pet_clair', 'autre', 'pehd', 'pet_fonce', 'environnement'] Configurations: BACKBONE resnet101 BACKBONE_SHAPES [[160 160] [ 80 80] [ 40 40] [ 20 20] [ 10 10]] BACKBONE_STRIDES [4, 8, 16, 32, 64] BATCH_SIZE 1 BBOX_STD_DEV [0.1 0.1 0.2 0.2] DETECTION_MAX_INSTANCES 100 DETECTION_MIN_CONFIDENCE 0.3 DETECTION_NMS_THRESHOLD 0.3 GPU_COUNT 1 IMAGES_PER_GPU 1 IMAGE_MAX_DIM 640 IMAGE_MIN_DIM 640 IMAGE_PADDING True 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 100 MEAN_PIXEL [123.7 116.8 103.9] MINI_MASK_SHAPE (56, 56) NAME learn_RUBBIA_REFUS_AMIENS_23 NUM_CLASSES 9 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 (16, 32, 64, 128, 256) 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 1000 TRAIN_ROIS_PER_IMAGE 200 USE_MINI_MASK True USE_RPN_ROIS True VALIDATION_STEPS 50 WEIGHT_DECAY 0.0001 model_param file didn't exist model_name : learn_RUBBIA_REFUS_AMIENS_23 model_type : mask_rcnn list file need : ['mask_model.h5'] file exist in s3 : ['mask_model.h5'] file manque in s3 : [] 2025-03-11 14:20:41.679784: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcublas.so.10 2025-03-11 14:20:41.873035: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7 local folder : /data/models_weight/learn_RUBBIA_REFUS_AMIENS_23 /data/models_weight/learn_RUBBIA_REFUS_AMIENS_23/mask_model.h5 size_local : 256009536 size in s3 : 256009536 create time local : 2021-08-09 09:43:22 create time in s3 : 2021-08-06 18:54:04 mask_model.h5 already exist and didn't need to update list_images length : 5 NEW PHOTO pour l'instant on ne peut pas sauvegarder la photo dans les tile Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -120.41875 max: 150.03750 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 6 time for calcul the mask position with numpy : 0.00012302398681640625 nb_pixel_total : 506 time to create 1 rle with old method : 0.001238107681274414 length of segment : 43 time for calcul the mask position with numpy : 6.008148193359375e-05 nb_pixel_total : 21 time to create 1 rle with old method : 9.083747863769531e-05 length of segment : 8 time for calcul the mask position with numpy : 5.030632019042969e-05 nb_pixel_total : 105 time to create 1 rle with old method : 0.0002720355987548828 length of segment : 23 time for calcul the mask position with numpy : 4.220008850097656e-05 nb_pixel_total : 26 time to create 1 rle with old method : 8.416175842285156e-05 length of segment : 15 time for calcul the mask position with numpy : 3.814697265625e-05 nb_pixel_total : 8 time to create 1 rle with old method : 5.91278076171875e-05 length of segment : 2 time for calcul the mask position with numpy : 3.981590270996094e-05 nb_pixel_total : 104 time to create 1 rle with old method : 0.00021958351135253906 length of segment : 26 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -117.97344 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 6 time for calcul the mask position with numpy : 6.604194641113281e-05 nb_pixel_total : 2477 time to create 1 rle with old method : 0.0031991004943847656 length of segment : 50 time for calcul the mask position with numpy : 5.6743621826171875e-05 nb_pixel_total : 2000 time to create 1 rle with old method : 0.0025849342346191406 length of segment : 39 time for calcul the mask position with numpy : 7.486343383789062e-05 nb_pixel_total : 2545 time to create 1 rle with old method : 0.0031735897064208984 length of segment : 94 time for calcul the mask position with numpy : 4.8160552978515625e-05 nb_pixel_total : 1186 time to create 1 rle with old method : 0.0015785694122314453 length of segment : 38 time for calcul the mask position with numpy : 4.38690185546875e-05 nb_pixel_total : 1146 time to create 1 rle with old method : 0.0015139579772949219 length of segment : 34 time for calcul the mask position with numpy : 7.295608520507812e-05 nb_pixel_total : 2703 time to create 1 rle with old method : 0.003298521041870117 length of segment : 101 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -113.16094 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 1 time for calcul the mask position with numpy : 8.749961853027344e-05 nb_pixel_total : 4377 time to create 1 rle with old method : 0.0051190853118896484 length of segment : 126 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -119.53984 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 5 time for calcul the mask position with numpy : 5.0067901611328125e-05 nb_pixel_total : 437 time to create 1 rle with old method : 0.0007925033569335938 length of segment : 20 time for calcul the mask position with numpy : 4.124641418457031e-05 nb_pixel_total : 257 time to create 1 rle with old method : 0.0004353523254394531 length of segment : 40 time for calcul the mask position with numpy : 3.504753112792969e-05 nb_pixel_total : 491 time to create 1 rle with old method : 0.0006833076477050781 length of segment : 22 time for calcul the mask position with numpy : 8.058547973632812e-05 nb_pixel_total : 2181 time to create 1 rle with old method : 0.0030031204223632812 length of segment : 106 time for calcul the mask position with numpy : 8.20159912109375e-05 nb_pixel_total : 2315 time to create 1 rle with old method : 0.0032405853271484375 length of segment : 104 Processing 1 images image shape: (400, 320, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 135.97891 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 8 time for calcul the mask position with numpy : 4.839897155761719e-05 nb_pixel_total : 474 time to create 1 rle with old method : 0.0007305145263671875 length of segment : 16 time for calcul the mask position with numpy : 3.814697265625e-05 nb_pixel_total : 277 time to create 1 rle with old method : 0.0003826618194580078 length of segment : 29 time for calcul the mask position with numpy : 5.555152893066406e-05 nb_pixel_total : 1069 time to create 1 rle with old method : 0.0013349056243896484 length of segment : 106 time for calcul the mask position with numpy : 4.0531158447265625e-05 nb_pixel_total : 481 time to create 1 rle with old method : 0.0007429122924804688 length of segment : 17 time for calcul the mask position with numpy : 0.0005695819854736328 nb_pixel_total : 45980 time to create 1 rle with old method : 0.08309364318847656 length of segment : 282 time for calcul the mask position with numpy : 0.0009367465972900391 nb_pixel_total : 68655 time to create 1 rle with old method : 0.07707023620605469 length of segment : 385 time for calcul the mask position with numpy : 7.724761962890625e-05 nb_pixel_total : 935 time to create 1 rle with old method : 0.0012781620025634766 length of segment : 30 time for calcul the mask position with numpy : 0.0006451606750488281 nb_pixel_total : 56288 time to create 1 rle with old method : 0.06274533271789551 length of segment : 356 Processing 1 images image shape: (400, 400, 3) min: 1.00000 max: 168.00000 molded_images shape: (1, 640, 640, 3) min: -102.02422 max: 41.71172 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 0 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 253.00000 molded_images shape: (1, 640, 640, 3) min: -119.96172 max: 133.62734 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 4 time for calcul the mask position with numpy : 6.556510925292969e-05 nb_pixel_total : 169 time to create 1 rle with old method : 0.0004322528839111328 length of segment : 17 time for calcul the mask position with numpy : 0.00012302398681640625 nb_pixel_total : 3003 time to create 1 rle with old method : 0.0038428306579589844 length of segment : 63 time for calcul the mask position with numpy : 4.2438507080078125e-05 nb_pixel_total : 262 time to create 1 rle with old method : 0.00037384033203125 length of segment : 21 time for calcul the mask position with numpy : 0.0001327991485595703 nb_pixel_total : 2990 time to create 1 rle with old method : 0.0037450790405273438 length of segment : 64 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -114.77813 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 5 time for calcul the mask position with numpy : 5.078315734863281e-05 nb_pixel_total : 485 time to create 1 rle with old method : 0.0009322166442871094 length of segment : 26 time for calcul the mask position with numpy : 4.100799560546875e-05 nb_pixel_total : 98 time to create 1 rle with old method : 0.00023865699768066406 length of segment : 9 time for calcul the mask position with numpy : 3.647804260253906e-05 nb_pixel_total : 439 time to create 1 rle with old method : 0.0006453990936279297 length of segment : 15 time for calcul the mask position with numpy : 3.838539123535156e-05 nb_pixel_total : 330 time to create 1 rle with old method : 0.0005085468292236328 length of segment : 23 time for calcul the mask position with numpy : 4.172325134277344e-05 nb_pixel_total : 348 time to create 1 rle with old method : 0.0005528926849365234 length of segment : 18 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -121.84453 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 8 time for calcul the mask position with numpy : 5.245208740234375e-05 nb_pixel_total : 317 time to create 1 rle with old method : 0.0004878044128417969 length of segment : 19 time for calcul the mask position with numpy : 9.036064147949219e-05 nb_pixel_total : 1086 time to create 1 rle with old method : 0.0014505386352539062 length of segment : 104 time for calcul the mask position with numpy : 6.67572021484375e-05 nb_pixel_total : 1240 time to create 1 rle with old method : 0.0016269683837890625 length of segment : 40 time for calcul the mask position with numpy : 5.793571472167969e-05 nb_pixel_total : 291 time to create 1 rle with old method : 0.0005991458892822266 length of segment : 25 time for calcul the mask position with numpy : 0.0005877017974853516 nb_pixel_total : 16273 time to create 1 rle with old method : 0.037925004959106445 length of segment : 285 time for calcul the mask position with numpy : 4.982948303222656e-05 nb_pixel_total : 66 time to create 1 rle with old method : 0.00011849403381347656 length of segment : 17 time for calcul the mask position with numpy : 3.528594970703125e-05 nb_pixel_total : 148 time to create 1 rle with old method : 0.0002315044403076172 length of segment : 17 time for calcul the mask position with numpy : 4.2438507080078125e-05 nb_pixel_total : 805 time to create 1 rle with old method : 0.001041412353515625 length of segment : 37 Processing 1 images image shape: (400, 320, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 9 time for calcul the mask position with numpy : 0.00012373924255371094 nb_pixel_total : 1780 time to create 1 rle with old method : 0.003244161605834961 length of segment : 114 time for calcul the mask position with numpy : 0.00021195411682128906 nb_pixel_total : 4685 time to create 1 rle with old method : 0.008010149002075195 length of segment : 238 time for calcul the mask position with numpy : 9.298324584960938e-05 nb_pixel_total : 1535 time to create 1 rle with old method : 0.002670764923095703 length of segment : 81 time for calcul the mask position with numpy : 5.936622619628906e-05 nb_pixel_total : 499 time to create 1 rle with old method : 0.0009107589721679688 length of segment : 32 time for calcul the mask position with numpy : 7.05718994140625e-05 nb_pixel_total : 827 time to create 1 rle with old method : 0.0014781951904296875 length of segment : 52 time for calcul the mask position with numpy : 0.0002307891845703125 nb_pixel_total : 4399 time to create 1 rle with old method : 0.005414247512817383 length of segment : 231 time for calcul the mask position with numpy : 9.560585021972656e-05 nb_pixel_total : 1507 time to create 1 rle with old method : 0.0019378662109375 length of segment : 131 time for calcul the mask position with numpy : 5.316734313964844e-05 nb_pixel_total : 347 time to create 1 rle with old method : 0.0005061626434326172 length of segment : 52 time for calcul the mask position with numpy : 0.00019812583923339844 nb_pixel_total : 4692 time to create 1 rle with old method : 0.0057888031005859375 length of segment : 219 Processing 1 images image shape: (280, 400, 3) min: 0.00000 max: 166.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 44.20781 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 1 time for calcul the mask position with numpy : 0.001132965087890625 nb_pixel_total : 105701 time to create 1 rle with old method : 0.13007140159606934 length of segment : 282 Processing 1 images image shape: (280, 400, 3) min: 0.00000 max: 217.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 104.96328 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 5 time for calcul the mask position with numpy : 4.9591064453125e-05 nb_pixel_total : 351 time to create 1 rle with old method : 0.0005037784576416016 length of segment : 24 time for calcul the mask position with numpy : 0.00010132789611816406 nb_pixel_total : 2540 time to create 1 rle with old method : 0.0031261444091796875 length of segment : 165 time for calcul the mask position with numpy : 6.4849853515625e-05 nb_pixel_total : 2394 time to create 1 rle with old method : 0.003017425537109375 length of segment : 85 time for calcul the mask position with numpy : 4.5299530029296875e-05 nb_pixel_total : 852 time to create 1 rle with old method : 0.0011551380157470703 length of segment : 58 time for calcul the mask position with numpy : 5.650520324707031e-05 nb_pixel_total : 1129 time to create 1 rle with old method : 0.0018281936645507812 length of segment : 35 Processing 1 images image shape: (280, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 139.88906 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 27 time for calcul the mask position with numpy : 6.723403930664062e-05 nb_pixel_total : 674 time to create 1 rle with old method : 0.000990152359008789 length of segment : 34 time for calcul the mask position with numpy : 9.369850158691406e-05 nb_pixel_total : 871 time to create 1 rle with old method : 0.001588582992553711 length of segment : 54 time for calcul the mask position with numpy : 6.175041198730469e-05 nb_pixel_total : 1010 time to create 1 rle with old method : 0.001405954360961914 length of segment : 26 time for calcul the mask position with numpy : 4.410743713378906e-05 nb_pixel_total : 207 time to create 1 rle with old method : 0.00034880638122558594 length of segment : 17 time for calcul the mask position with numpy : 6.127357482910156e-05 nb_pixel_total : 1036 time to create 1 rle with old method : 0.0013422966003417969 length of segment : 45 time for calcul the mask position with numpy : 8.058547973632812e-05 nb_pixel_total : 1850 time to create 1 rle with old method : 0.002345561981201172 length of segment : 65 time for calcul the mask position with numpy : 4.6253204345703125e-05 nb_pixel_total : 290 time to create 1 rle with old method : 0.00043773651123046875 length of segment : 22 time for calcul the mask position with numpy : 4.00543212890625e-05 nb_pixel_total : 87 time to create 1 rle with old method : 0.00018358230590820312 length of segment : 6 time for calcul the mask position with numpy : 4.7206878662109375e-05 nb_pixel_total : 368 time to create 1 rle with old method : 0.0005345344543457031 length of segment : 26 time for calcul the mask position with numpy : 5.221366882324219e-05 nb_pixel_total : 503 time to create 1 rle with old method : 0.0007729530334472656 length of segment : 38 time for calcul the mask position with numpy : 6.270408630371094e-05 nb_pixel_total : 168 time to create 1 rle with old method : 0.00041961669921875 length of segment : 18 time for calcul the mask position with numpy : 4.076957702636719e-05 nb_pixel_total : 149 time to create 1 rle with old method : 0.000247955322265625 length of segment : 12 time for calcul the mask position with numpy : 4.553794860839844e-05 nb_pixel_total : 351 time to create 1 rle with old method : 0.0004994869232177734 length of segment : 30 time for calcul the mask position with numpy : 3.838539123535156e-05 nb_pixel_total : 78 time to create 1 rle with old method : 0.00013589859008789062 length of segment : 12 time for calcul the mask position with numpy : 3.147125244140625e-05 nb_pixel_total : 118 time to create 1 rle with old method : 0.00020074844360351562 length of segment : 14 time for calcul the mask position with numpy : 4.363059997558594e-05 nb_pixel_total : 453 time to create 1 rle with old method : 0.0006418228149414062 length of segment : 33 time for calcul the mask position with numpy : 4.00543212890625e-05 nb_pixel_total : 130 time to create 1 rle with old method : 0.0002224445343017578 length of segment : 16 time for calcul the mask position with numpy : 6.866455078125e-05 nb_pixel_total : 225 time to create 1 rle with old method : 0.0003628730773925781 length of segment : 20 time for calcul the mask position with numpy : 3.981590270996094e-05 nb_pixel_total : 293 time to create 1 rle with old method : 0.00042819976806640625 length of segment : 19 time for calcul the mask position with numpy : 4.696846008300781e-05 nb_pixel_total : 413 time to create 1 rle with old method : 0.0006067752838134766 length of segment : 27 time for calcul the mask position with numpy : 9.846687316894531e-05 nb_pixel_total : 1797 time to create 1 rle with old method : 0.003289461135864258 length of segment : 65 time for calcul the mask position with numpy : 5.412101745605469e-05 nb_pixel_total : 173 time to create 1 rle with old method : 0.0003943443298339844 length of segment : 15 time for calcul the mask position with numpy : 5.221366882324219e-05 nb_pixel_total : 160 time to create 1 rle with old method : 0.0003681182861328125 length of segment : 12 time for calcul the mask position with numpy : 6.842613220214844e-05 nb_pixel_total : 250 time to create 1 rle with old method : 0.0004611015319824219 length of segment : 43 time for calcul the mask position with numpy : 7.200241088867188e-05 nb_pixel_total : 304 time to create 1 rle with old method : 0.0006842613220214844 length of segment : 42 time for calcul the mask position with numpy : 6.103515625e-05 nb_pixel_total : 235 time to create 1 rle with old method : 0.0009517669677734375 length of segment : 27 time for calcul the mask position with numpy : 6.508827209472656e-05 nb_pixel_total : 336 time to create 1 rle with old method : 0.0006940364837646484 length of segment : 20 Processing 1 images image shape: (280, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 150.20156 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 5 time for calcul the mask position with numpy : 4.9114227294921875e-05 nb_pixel_total : 246 time to create 1 rle with old method : 0.00036334991455078125 length of segment : 25 time for calcul the mask position with numpy : 7.700920104980469e-05 nb_pixel_total : 1348 time to create 1 rle with old method : 0.0018880367279052734 length of segment : 73 time for calcul the mask position with numpy : 0.00010156631469726562 nb_pixel_total : 2272 time to create 1 rle with old method : 0.003122091293334961 length of segment : 93 time for calcul the mask position with numpy : 7.462501525878906e-05 nb_pixel_total : 895 time to create 1 rle with old method : 0.0012536048889160156 length of segment : 53 time for calcul the mask position with numpy : 0.00017642974853515625 nb_pixel_total : 7563 time to create 1 rle with old method : 0.009476661682128906 length of segment : 147 Processing 1 images image shape: (280, 320, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -123.70000 max: 131.97500 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 3 time for calcul the mask position with numpy : 0.0007100105285644531 nb_pixel_total : 37241 time to create 1 rle with old method : 0.0470280647277832 length of segment : 243 time for calcul the mask position with numpy : 0.00013184547424316406 nb_pixel_total : 1 time to create 1 rle with old method : 3.457069396972656e-05 length of segment : 1 time for calcul the mask position with numpy : 5.817413330078125e-05 nb_pixel_total : 302 time to create 1 rle with old method : 0.0004892349243164062 length of segment : 71 NEW PHOTO pour l'instant on ne peut pas sauvegarder la photo dans les tile Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -118.14141 max: 147.78750 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 6 time for calcul the mask position with numpy : 7.724761962890625e-05 nb_pixel_total : 626 time to create 1 rle with old method : 0.0013136863708496094 length of segment : 43 time for calcul the mask position with numpy : 5.054473876953125e-05 nb_pixel_total : 8 time to create 1 rle with old method : 6.532669067382812e-05 length of segment : 1 time for calcul the mask position with numpy : 4.172325134277344e-05 nb_pixel_total : 31 time to create 1 rle with old method : 9.393692016601562e-05 length of segment : 9 time for calcul the mask position with numpy : 4.315376281738281e-05 nb_pixel_total : 90 time to create 1 rle with old method : 0.00021719932556152344 length of segment : 20 time for calcul the mask position with numpy : 5.364418029785156e-05 nb_pixel_total : 45 time to create 1 rle with old method : 0.00011587142944335938 length of segment : 17 time for calcul the mask position with numpy : 5.984306335449219e-05 nb_pixel_total : 702 time to create 1 rle with old method : 0.0013353824615478516 length of segment : 51 Processing 1 images image shape: (400, 400, 3) min: 0.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -118.30938 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 5 time for calcul the mask position with numpy : 7.05718994140625e-05 nb_pixel_total : 2274 time to create 1 rle with old method : 0.0029137134552001953 length of segment : 46 time for calcul the mask position with numpy : 7.82012939453125e-05 nb_pixel_total : 1328 time to create 1 rle with old method : 0.0017321109771728516 length of segment : 30 length of segment : 0 time for calcul the mask position with numpy : 9.393692016601562e-05 nb_pixel_total : 2266 time to create 1 rle with old method : 0.0028629302978515625 length of segment : 95 time for calcul the mask position with numpy : 0.00011277198791503906 nb_pixel_total : 2244 time to create 1 rle with old method : 0.0028374195098876953 length of segment : 95 Processing 1 images image shape: (400, 400, 3) min: 2.00000 max: 255.00000 molded_images shape: (1, 640, 640, 3) min: -111.57891 max: 151.10000 image_metas shape: (1, 17) min: 0.00000 max: 640.00000 nb d'objets trouves : 3 time for calcul the mask position with numpy : 8.368492126464844e-05 nb_pixel_total : 2661 time to create 1 rle with old method : 0.0034008026123046875 length of segment : 43 time for calcul the mask position with numpy : 0.0001430511474609375 nb_pixel_total : 3973 time to create 1 rle with old method : 0.0049402713775634766 length of segment : 140 time for calcul the mask position with numpy : 0.0001289844512939453 nb_pixel_total : 4107 time to create 1 rle with old method : 0.005037784576416016 2025-03-11 14:20:49.439875: E tensorflow/stream_executor/cuda/cuda_event.cc:29] Error polling for event status: failed to query event: CUDA_ERROR_ILLEGAL_ADDRESS: an illegal memory access was encountered 2025-03-11 14:20:49.439932: F tensorflow/core/common_runtime/gpu/gpu_event_mgr.cc:273] Unexpected Event status: 1 max_time_sub_proc : 3600 Catched exception ! Connect or reconnect ! in case -12 caffe_path_current : About to save ! 2 After save, about to update current ! 16.63user 12.49system 1:00:06elapsed 0%CPU (0avgtext+0avgdata 2522708maxresident)k 616776inputs+2624outputs (10529major+703992minor)pagefaults 0swaps