2022/09/15 17:19:22 - mmengine - INFO - ------------------------------------------------------------ System environment: sys.platform: linux Python: 3.7.13 (default, Mar 29 2022, 02:18:16) [GCC 7.5.0] CUDA available: True numpy_random_seed: 1113651553 GPU 0,1,2,3,4,5,6,7: NVIDIA A100-SXM4-80GB CUDA_HOME: /mnt/cache/share/cuda-11.1 NVCC: Cuda compilation tools, release 11.1, V11.1.74 GCC: gcc (GCC) 5.4.0 PyTorch: 1.9.0+cu111 PyTorch compiling details: PyTorch built with: - GCC 7.3 - C++ Version: 201402 - Intel(R) Math Kernel Library Version 2020.0.0 Product Build 20191122 for Intel(R) 64 architecture applications - Intel(R) MKL-DNN v2.1.2 (Git Hash 98be7e8afa711dc9b66c8ff3504129cb82013cdb) - OpenMP 201511 (a.k.a. OpenMP 4.5) - NNPACK is enabled - CPU capability usage: AVX2 - CUDA Runtime 11.1 - NVCC architecture flags: -gencode;arch=compute_37,code=sm_37;-gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86 - CuDNN 8.0.5 - Magma 2.5.2 - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.1, CUDNN_VERSION=8.0.5, CXX_COMPILER=/opt/rh/devtoolset-7/root/usr/bin/c++, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -fopenmp -DNDEBUG -DUSE_KINETO -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-unused-local-typedefs -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-stringop-overflow -Wno-psabi -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=1.9.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, TorchVision: 0.10.0+cu111 OpenCV: 4.5.4 MMEngine: 0.1.0 Runtime environment: cudnn_benchmark: True mp_cfg: {'mp_start_method': 'fork', 'opencv_num_threads': 0} dist_cfg: {'backend': 'nccl'} seed: None Distributed launcher: slurm Distributed training: True GPU number: 8 ------------------------------------------------------------ 2022/09/15 17:19:24 - mmengine - INFO - Config: mj_rec_data_root = 'data/rec/Syn90k/' mj_rec_train = dict( type='OCRDataset', data_root='data/rec/Syn90k/', data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'), ann_file='train_labels.json', test_mode=False, pipeline=None) mj_sub_rec_train = dict( type='OCRDataset', data_root='data/rec/Syn90k/', data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'), ann_file='subset_train_labels.json', test_mode=False, pipeline=None) st_data_root = 'data/rec/SynthText/' st_rec_train = dict( type='OCRDataset', data_root='data/rec/SynthText/', data_prefix=dict(img_path='synthtext/SynthText_patch_horizontal'), ann_file='train_labels.json', test_mode=False, pipeline=None) st_an_rec_train = dict( type='OCRDataset', data_root='data/rec/SynthText/', data_prefix=dict(img_path='synthtext/SynthText_patch_horizontal'), ann_file='alphanumeric_train_labels.json', test_mode=False, pipeline=None) st_sub_rec_train = dict( type='OCRDataset', data_root='data/rec/SynthText/', data_prefix=dict(img_path='synthtext/SynthText_patch_horizontal'), ann_file='subset_train_labels.json', test_mode=False, pipeline=None) st_add_rec_data_root = 'data/rec/synthtext_add/' st_add_rec_train = dict( type='OCRDataset', data_root='data/rec/synthtext_add/', ann_file='train_labels.json', test_mode=False, pipeline=None) cocov1_rec_train_data_root = 'data/rec/coco_text_v1' cocov1_rec_train = dict( type='OCRDataset', data_root='data/rec/coco_text_v1', ann_file='train_labels.json', test_mode=False, pipeline=None) cute80_rec_data_root = 'data/rec/ct80/' cute80_rec_test = dict( type='OCRDataset', data_root='data/rec/ct80/', ann_file='test_labels.json', test_mode=True, pipeline=None) iiit5k_rec_data_root = 'data/rec/IIIT5K/' iiit5k_rec_train = dict( type='OCRDataset', data_root='data/rec/IIIT5K/', ann_file='train_labels.json', test_mode=False, pipeline=None) iiit5k_rec_test = dict( type='OCRDataset', data_root='data/rec/IIIT5K/', ann_file='test_labels.json', test_mode=True, pipeline=None) svt_rec_data_root = 'data/rec/svt/' svt_rec_test = dict( type='OCRDataset', data_root='data/rec/svt/', ann_file='test_labels.json', test_mode=True, pipeline=None) svtp_rec_data_root = 'data/rec/svtp/' svtp_rec_test = dict( type='OCRDataset', data_root='data/rec/svtp/', ann_file='test_labels.json', test_mode=True, pipeline=None) ic11_rec_data_root = 'data/rec/icdar_2011/' ic11_rec_train = dict( type='OCRDataset', data_root='data/rec/icdar_2011/', ann_file='train_labels.json', test_mode=False, pipeline=None) ic13_rec_data_root = 'data/rec/icdar_2013/' ic13_rec_train = dict( type='OCRDataset', data_root='data/rec/icdar_2013/', ann_file='train_labels.json', test_mode=False, pipeline=None) ic13_rec_test = dict( type='OCRDataset', data_root='data/rec/icdar_2013/', ann_file='test_labels.json', test_mode=True, pipeline=None) ic15_rec_data_root = 'data/rec/icdar_2015/' ic15_rec_train = dict( type='OCRDataset', data_root='data/rec/icdar_2015/', ann_file='train_labels.json', test_mode=False, pipeline=None) ic15_rec_test = dict( type='OCRDataset', data_root='data/rec/icdar_2015/', ann_file='test_labels.json', test_mode=True, pipeline=None) default_scope = 'mmocr' env_cfg = dict( cudnn_benchmark=True, mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0), dist_cfg=dict(backend='nccl')) randomness = dict(seed=None) default_hooks = dict( timer=dict(type='IterTimerHook'), logger=dict(type='LoggerHook', interval=100), param_scheduler=dict(type='ParamSchedulerHook'), checkpoint=dict( type='CheckpointHook', interval=1, out_dir='sproject:s3://1.0.0rc0_recog_retest'), sampler_seed=dict(type='DistSamplerSeedHook'), sync_buffer=dict(type='SyncBuffersHook'), visualization=dict( type='VisualizationHook', interval=1, enable=False, show=False, draw_gt=False, draw_pred=False)) log_level = 'INFO' log_processor = dict(type='LogProcessor', window_size=10, by_epoch=True) load_from = None resume = False val_evaluator = dict( type='MultiDatasetsEvaluator', metrics=[dict(type='WordMetric', mode=['ignore_case_symbol'])], dataset_prefixes=['CUTE80', 'IIIT5K', 'SVT', 'SVTP', 'IC13', 'IC15']) test_evaluator = dict( type='MultiDatasetsEvaluator', metrics=[dict(type='WordMetric', mode=['ignore_case_symbol'])], dataset_prefixes=['CUTE80', 'IIIT5K', 'SVT', 'SVTP', 'IC13', 'IC15']) vis_backends = [dict(type='LocalVisBackend')] visualizer = dict( type='TextRecogLocalVisualizer', name='visualizer', vis_backends=[dict(type='LocalVisBackend')]) optim_wrapper = dict( type='OptimWrapper', optimizer=dict(type='Adam', lr=0.001)) train_cfg = dict(type='EpochBasedTrainLoop', max_epochs=5, val_interval=1) val_cfg = dict(type='ValLoop') test_cfg = dict(type='TestLoop') param_scheduler = [dict(type='MultiStepLR', milestones=[3, 4], end=5)] dictionary = dict( type='Dictionary', dict_file= 'configs/textrecog/sar/../../../dicts/english_digits_symbols.txt', with_start=True, with_end=True, same_start_end=True, with_padding=True, with_unknown=True) model = dict( type='SARNet', data_preprocessor=dict( type='TextRecogDataPreprocessor', mean=[127, 127, 127], std=[127, 127, 127]), backbone=dict(type='ResNet31OCR'), encoder=dict( type='SAREncoder', enc_bi_rnn=False, enc_do_rnn=0.1, enc_gru=False), decoder=dict( type='ParallelSARDecoder', enc_bi_rnn=False, dec_bi_rnn=False, dec_do_rnn=0, dec_gru=False, pred_dropout=0.1, d_k=512, pred_concat=True, postprocessor=dict(type='AttentionPostprocessor'), module_loss=dict( type='CEModuleLoss', ignore_first_char=True, reduction='mean'), dictionary=dict( type='Dictionary', dict_file= 'configs/textrecog/sar/../../../dicts/english_digits_symbols.txt', with_start=True, with_end=True, same_start_end=True, with_padding=True, with_unknown=True), max_seq_len=30)) file_client_args = dict(backend='disk') train_pipeline = [ dict( type='LoadImageFromFile', file_client_args=dict(backend='disk'), ignore_empty=True, min_size=2), dict(type='LoadOCRAnnotations', with_text=True), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ] test_pipeline = [ dict(type='LoadImageFromFile', file_client_args=dict(backend='disk')), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict(type='LoadOCRAnnotations', with_text=True), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ] train_list = [ dict( type='RepeatDataset', dataset=dict( type='ConcatDataset', datasets=[ dict( type='OCRDataset', data_root='data/rec/icdar_2011/', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root='data/rec/icdar_2013/', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root='data/rec/icdar_2015/', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root='data/rec/coco_text_v1', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root='data/rec/IIIT5K/', ann_file='train_labels.json', test_mode=False, pipeline=None) ], pipeline=[ dict( type='LoadImageFromFile', file_client_args=dict(backend='disk'), ignore_empty=True, min_size=2), dict(type='LoadOCRAnnotations', with_text=True), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ]), times=20), dict( type='ConcatDataset', datasets=[ dict( type='OCRDataset', data_root='data/rec/Syn90k/', data_prefix=dict(img_path='mnt/ramdisk/max/90kDICT32px'), ann_file='subset_train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root='data/rec/SynthText/', data_prefix=dict( img_path='synthtext/SynthText_patch_horizontal'), ann_file='subset_train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root='data/rec/synthtext_add/', ann_file='train_labels.json', test_mode=False, pipeline=None) ], pipeline=[ dict( type='LoadImageFromFile', file_client_args=dict(backend='disk'), ignore_empty=True, min_size=2), dict(type='LoadOCRAnnotations', with_text=True), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ]) ] test_list = [ dict( type='OCRDataset', data_root='data/rec/ct80/', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='data/rec/IIIT5K/', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='data/rec/svt/', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='data/rec/svtp/', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='data/rec/icdar_2013/', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='data/rec/icdar_2015/', ann_file='test_labels.json', test_mode=True, pipeline=None) ] train_dataloader = dict( batch_size=384, num_workers=24, persistent_workers=True, sampler=dict(type='DefaultSampler', shuffle=True), dataset=dict( type='ConcatDataset', datasets=[ dict( type='RepeatDataset', dataset=dict( type='ConcatDataset', datasets=[ dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2011', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2013', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2015', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/coco_text_v1', ann_file='train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/IIIT5K', ann_file='train_labels.json', test_mode=False, pipeline=None) ], pipeline=[ dict( type='LoadImageFromFile', file_client_args=dict(backend='petrel'), ignore_empty=True, min_size=2), dict(type='LoadOCRAnnotations', with_text=True), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ]), times=20), dict( type='ConcatDataset', datasets=[ dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/Syn90k', data_prefix=dict( img_path='mnt/ramdisk/max/90kDICT32px'), ann_file='subset_train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/SynthText', data_prefix=dict( img_path='synthtext/SynthText_patch_horizontal'), ann_file='subset_train_labels.json', test_mode=False, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/synthtext_add', ann_file='train_labels.json', test_mode=False, pipeline=None) ], pipeline=[ dict( type='LoadImageFromFile', file_client_args=dict(backend='petrel'), ignore_empty=True, min_size=2), dict(type='LoadOCRAnnotations', with_text=True), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ]) ], verify_meta=False)) test_dataloader = dict( batch_size=1, num_workers=4, persistent_workers=True, drop_last=False, sampler=dict(type='DefaultSampler', shuffle=False), dataset=dict( type='ConcatDataset', datasets=[ dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/ct80', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/IIIT5K', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/svt', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/svtp', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2013', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2015', ann_file='test_labels.json', test_mode=True, pipeline=None) ], pipeline=[ dict( type='LoadImageFromFile', file_client_args=dict(backend='petrel')), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict(type='LoadOCRAnnotations', with_text=True), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ])) val_dataloader = dict( batch_size=1, num_workers=4, persistent_workers=True, drop_last=False, sampler=dict(type='DefaultSampler', shuffle=False), dataset=dict( type='ConcatDataset', datasets=[ dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/ct80', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/IIIT5K', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/svt', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root='openmmlab:s3://openmmlab/datasets/ocr/recog/svtp', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2013', ann_file='test_labels.json', test_mode=True, pipeline=None), dict( type='OCRDataset', data_root= 'openmmlab:s3://openmmlab/datasets/ocr/recog/icdar_2015', ann_file='test_labels.json', test_mode=True, pipeline=None) ], pipeline=[ dict( type='LoadImageFromFile', file_client_args=dict(backend='petrel')), dict( type='RescaleToHeight', height=48, min_width=48, max_width=160, width_divisor=4), dict(type='PadToWidth', width=160), dict(type='LoadOCRAnnotations', with_text=True), dict( type='PackTextRecogInputs', meta_keys=('img_path', 'ori_shape', 'img_shape', 'valid_ratio')) ])) launcher = 'slurm' work_dir = './work_dirs/sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real' Name of parameter - Initialization information backbone.conv1_1.weight - torch.Size([64, 3, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.conv1_1.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.bn1_1.weight - torch.Size([64]): UniformInit: a=0, b=1, bias=0 backbone.bn1_1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of SARNet backbone.conv1_2.weight - torch.Size([128, 64, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.conv1_2.bias - torch.Size([128]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.bn1_2.weight - torch.Size([128]): UniformInit: a=0, b=1, bias=0 backbone.bn1_2.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of SARNet backbone.block2.0.conv1.weight - torch.Size([256, 128, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block2.0.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block2.0.bn1.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block2.0.bn1.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block2.0.bn2.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block2.0.bn2.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block2.0.downsample.0.weight - torch.Size([256, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block2.0.downsample.1.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block2.0.downsample.1.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.conv2.bias - torch.Size([256]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.bn2.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.bn2.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block3.0.conv1.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block3.0.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block3.0.bn1.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block3.0.bn1.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block3.0.bn2.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block3.0.bn2.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block3.1.conv1.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block3.1.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block3.1.bn1.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block3.1.bn1.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block3.1.bn2.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.block3.1.bn2.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.conv3.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.conv3.bias - torch.Size([256]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.bn3.weight - torch.Size([256]): UniformInit: a=0, b=1, bias=0 backbone.bn3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.0.conv1.weight - torch.Size([512, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.0.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.0.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.0.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.0.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.0.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.0.downsample.0.weight - torch.Size([512, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.0.downsample.1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.0.downsample.1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.1.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.1.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.1.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.1.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.1.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.1.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.2.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.2.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.2.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.2.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.2.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.2.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.3.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.3.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.3.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.3.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.3.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.3.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.4.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.4.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block4.4.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.4.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block4.4.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block4.4.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.conv4.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.conv4.bias - torch.Size([512]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.bn4.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.bn4.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block5.0.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block5.0.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block5.0.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block5.0.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block5.0.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block5.0.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block5.1.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block5.1.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block5.1.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block5.1.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block5.1.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block5.1.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block5.2.conv1.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block5.2.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.block5.2.bn1.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block5.2.bn1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.block5.2.bn2.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.block5.2.bn2.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet backbone.conv5.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.conv5.bias - torch.Size([512]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.bn5.weight - torch.Size([512]): UniformInit: a=0, b=1, bias=0 backbone.bn5.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.weight_ih_l0 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.weight_hh_l0 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.bias_ih_l0 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.bias_hh_l0 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.weight_ih_l1 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.weight_hh_l1 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.bias_ih_l1 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet encoder.rnn_encoder.bias_hh_l1 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet encoder.linear.weight - torch.Size([512, 512]): The value is the same before and after calling `init_weights` of SARNet encoder.linear.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet decoder.conv1x1_1.weight - torch.Size([512, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.conv1x1_1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet decoder.conv3x3_1.weight - torch.Size([512, 512, 3, 3]): The value is the same before and after calling `init_weights` of SARNet decoder.conv3x3_1.bias - torch.Size([512]): The value is the same before and after calling `init_weights` of SARNet decoder.conv1x1_2.weight - torch.Size([1, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.conv1x1_2.bias - torch.Size([1]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.weight_ih_l0 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.weight_hh_l0 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.bias_ih_l0 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.bias_hh_l0 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.weight_ih_l1 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.weight_hh_l1 - torch.Size([2048, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.bias_ih_l1 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet decoder.rnn_decoder.bias_hh_l1 - torch.Size([2048]): The value is the same before and after calling `init_weights` of SARNet decoder.embedding.weight - torch.Size([93, 512]): The value is the same before and after calling `init_weights` of SARNet decoder.prediction.weight - torch.Size([93, 1536]): The value is the same before and after calling `init_weights` of SARNet decoder.prediction.bias - torch.Size([93]): The value is the same before and after calling `init_weights` of SARNet 2022/09/15 17:22:06 - mmengine - INFO - Checkpoints will be saved to sproject:s3://1.0.0rc0_recog_retest/sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real by PetrelBackend. 2022/09/15 17:46:33 - mmengine - INFO - Epoch(train) [1][100/2304] lr: 1.0000e-03 eta: 1 day, 22:32:46 time: 1.4619 data_time: 0.1717 memory: 44223 loss_ce: 2.5693 loss: 2.5693 2022/09/15 17:48:43 - mmengine - INFO - Epoch(train) [1][200/2304] lr: 1.0000e-03 eta: 1 day, 1:07:16 time: 1.6786 data_time: 0.2586 memory: 44223 loss_ce: 2.3220 loss: 2.3220 2022/09/15 17:50:49 - mmengine - INFO - Epoch(train) [1][300/2304] lr: 1.0000e-03 eta: 17:54:10 time: 1.8740 data_time: 0.1738 memory: 44223 loss_ce: 1.8770 loss: 1.8770 2022/09/15 17:52:50 - mmengine - INFO - Epoch(train) [1][400/2304] lr: 1.0000e-03 eta: 14:14:39 time: 0.9706 data_time: 0.1251 memory: 44223 loss_ce: 1.0869 loss: 1.0869 2022/09/15 17:54:55 - mmengine - INFO - Epoch(train) [1][500/2304] lr: 1.0000e-03 eta: 12:03:21 time: 0.8613 data_time: 0.0062 memory: 44223 loss_ce: 0.8470 loss: 0.8470 2022/09/15 17:56:59 - mmengine - INFO - Epoch(train) [1][600/2304] lr: 1.0000e-03 eta: 10:35:04 time: 0.8406 data_time: 0.0218 memory: 44223 loss_ce: 0.7175 loss: 0.7175 2022/09/15 17:59:11 - mmengine - INFO - Epoch(train) [1][700/2304] lr: 1.0000e-03 eta: 9:33:17 time: 1.4376 data_time: 0.1674 memory: 44223 loss_ce: 0.6416 loss: 0.6416 2022/09/15 18:01:21 - mmengine - INFO - Epoch(train) [1][800/2304] lr: 1.0000e-03 eta: 8:46:04 time: 1.7637 data_time: 0.3055 memory: 44223 loss_ce: 0.6122 loss: 0.6122 2022/09/15 18:03:28 - mmengine - INFO - Epoch(train) [1][900/2304] lr: 1.0000e-03 eta: 8:08:14 time: 1.8276 data_time: 0.1682 memory: 44223 loss_ce: 0.5719 loss: 0.5719 2022/09/15 18:05:29 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 18:05:29 - mmengine - INFO - Epoch(train) [1][1000/2304] lr: 1.0000e-03 eta: 7:36:27 time: 0.9442 data_time: 0.1239 memory: 44223 loss_ce: 0.5526 loss: 0.5526 2022/09/15 18:07:36 - mmengine - INFO - Epoch(train) [1][1100/2304] lr: 1.0000e-03 eta: 7:11:08 time: 0.8650 data_time: 0.0062 memory: 44223 loss_ce: 0.5557 loss: 0.5557 2022/09/15 18:09:42 - mmengine - INFO - Epoch(train) [1][1200/2304] lr: 1.0000e-03 eta: 6:49:23 time: 0.8658 data_time: 0.0186 memory: 44223 loss_ce: 0.5268 loss: 0.5268 2022/09/15 18:11:51 - mmengine - INFO - Epoch(train) [1][1300/2304] lr: 1.0000e-03 eta: 6:31:11 time: 1.4130 data_time: 0.1740 memory: 44223 loss_ce: 0.5009 loss: 0.5009 2022/09/15 18:13:57 - mmengine - INFO - Epoch(train) [1][1400/2304] lr: 1.0000e-03 eta: 6:14:52 time: 1.6637 data_time: 0.2729 memory: 44223 loss_ce: 0.5180 loss: 0.5180 2022/09/15 18:16:01 - mmengine - INFO - Epoch(train) [1][1500/2304] lr: 1.0000e-03 eta: 6:00:11 time: 1.7682 data_time: 0.1632 memory: 44223 loss_ce: 0.5128 loss: 0.5128 2022/09/15 18:18:01 - mmengine - INFO - Epoch(train) [1][1600/2304] lr: 1.0000e-03 eta: 5:46:45 time: 0.9730 data_time: 0.1255 memory: 44223 loss_ce: 0.5164 loss: 0.5164 2022/09/15 18:20:06 - mmengine - INFO - Epoch(train) [1][1700/2304] lr: 1.0000e-03 eta: 5:35:02 time: 0.8379 data_time: 0.0067 memory: 44223 loss_ce: 0.4929 loss: 0.4929 2022/09/15 18:22:10 - mmengine - INFO - Epoch(train) [1][1800/2304] lr: 1.0000e-03 eta: 5:24:20 time: 0.8626 data_time: 0.0192 memory: 44223 loss_ce: 0.4833 loss: 0.4833 2022/09/15 18:24:19 - mmengine - INFO - Epoch(train) [1][1900/2304] lr: 1.0000e-03 eta: 5:15:03 time: 1.5075 data_time: 0.1823 memory: 44223 loss_ce: 0.4516 loss: 0.4516 2022/09/15 18:26:24 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 18:26:24 - mmengine - INFO - Epoch(train) [1][2000/2304] lr: 1.0000e-03 eta: 5:06:05 time: 1.6670 data_time: 0.2762 memory: 44223 loss_ce: 0.4623 loss: 0.4623 2022/09/15 18:28:28 - mmengine - INFO - Epoch(train) [1][2100/2304] lr: 1.0000e-03 eta: 4:57:44 time: 1.7214 data_time: 0.1455 memory: 44223 loss_ce: 0.4658 loss: 0.4658 2022/09/15 18:30:29 - mmengine - INFO - Epoch(train) [1][2200/2304] lr: 1.0000e-03 eta: 4:49:41 time: 0.9967 data_time: 0.1193 memory: 44223 loss_ce: 0.4402 loss: 0.4402 2022/09/15 18:32:28 - mmengine - INFO - Epoch(train) [1][2300/2304] lr: 1.0000e-03 eta: 4:42:07 time: 0.8306 data_time: 0.0069 memory: 44223 loss_ce: 0.4608 loss: 0.4608 2022/09/15 18:32:50 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 18:32:50 - mmengine - INFO - Saving checkpoint at 1 epochs 2022/09/15 18:38:08 - mmengine - INFO - Epoch(val) [1][100/959] eta: 0:01:02 time: 0.0727 data_time: 0.0030 memory: 44223 2022/09/15 18:38:15 - mmengine - INFO - Epoch(val) [1][200/959] eta: 0:01:00 time: 0.0802 data_time: 0.0070 memory: 1130 2022/09/15 18:38:22 - mmengine - INFO - Epoch(val) [1][300/959] eta: 0:00:50 time: 0.0764 data_time: 0.0022 memory: 1130 2022/09/15 18:38:29 - mmengine - INFO - Epoch(val) [1][400/959] eta: 0:00:41 time: 0.0741 data_time: 0.0052 memory: 1130 2022/09/15 18:38:36 - mmengine - INFO - Epoch(val) [1][500/959] eta: 0:00:32 time: 0.0699 data_time: 0.0008 memory: 1130 2022/09/15 18:38:43 - mmengine - INFO - Epoch(val) [1][600/959] eta: 0:00:25 time: 0.0715 data_time: 0.0021 memory: 1130 2022/09/15 18:38:51 - mmengine - INFO - Epoch(val) [1][700/959] eta: 0:00:16 time: 0.0652 data_time: 0.0007 memory: 1130 2022/09/15 18:38:57 - mmengine - INFO - Epoch(val) [1][800/959] eta: 0:00:10 time: 0.0650 data_time: 0.0008 memory: 1130 2022/09/15 18:39:04 - mmengine - INFO - Epoch(val) [1][900/959] eta: 0:00:03 time: 0.0656 data_time: 0.0008 memory: 1130 2022/09/15 18:39:09 - mmengine - INFO - Epoch(val) [1][959/959] CUTE80/recog/word_acc_ignore_case_symbol: 0.8542 IIIT5K/recog/word_acc_ignore_case_symbol: 0.9203 SVT/recog/word_acc_ignore_case_symbol: 0.8470 SVTP/recog/word_acc_ignore_case_symbol: 0.7674 IC13/recog/word_acc_ignore_case_symbol: 0.9103 IC15/recog/word_acc_ignore_case_symbol: 0.7130 2022/09/15 18:41:19 - mmengine - INFO - Epoch(train) [2][100/2304] lr: 1.0000e-03 eta: 4:35:03 time: 1.3406 data_time: 0.3125 memory: 44223 loss_ce: 0.4293 loss: 0.4293 2022/09/15 18:43:22 - mmengine - INFO - Epoch(train) [2][200/2304] lr: 1.0000e-03 eta: 4:28:34 time: 1.4250 data_time: 0.2800 memory: 44222 loss_ce: 0.4291 loss: 0.4291 2022/09/15 18:45:24 - mmengine - INFO - Epoch(train) [2][300/2304] lr: 1.0000e-03 eta: 4:22:20 time: 1.2673 data_time: 0.0766 memory: 44222 loss_ce: 0.4227 loss: 0.4227 2022/09/15 18:47:26 - mmengine - INFO - Epoch(train) [2][400/2304] lr: 1.0000e-03 eta: 4:16:28 time: 1.0636 data_time: 0.0058 memory: 44222 loss_ce: 0.4263 loss: 0.4263 2022/09/15 18:49:26 - mmengine - INFO - Epoch(train) [2][500/2304] lr: 1.0000e-03 eta: 4:10:44 time: 1.0190 data_time: 0.0206 memory: 44222 loss_ce: 0.4281 loss: 0.4281 2022/09/15 18:51:29 - mmengine - INFO - Epoch(train) [2][600/2304] lr: 1.0000e-03 eta: 4:05:23 time: 0.9453 data_time: 0.0170 memory: 44222 loss_ce: 0.4301 loss: 0.4301 2022/09/15 18:53:27 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 18:53:35 - mmengine - INFO - Epoch(train) [2][700/2304] lr: 1.0000e-03 eta: 4:00:25 time: 1.4002 data_time: 0.3188 memory: 44222 loss_ce: 0.4074 loss: 0.4074 2022/09/15 18:55:39 - mmengine - INFO - Epoch(train) [2][800/2304] lr: 1.0000e-03 eta: 3:55:33 time: 1.4181 data_time: 0.3153 memory: 44222 loss_ce: 0.4158 loss: 0.4158 2022/09/15 18:57:45 - mmengine - INFO - Epoch(train) [2][900/2304] lr: 1.0000e-03 eta: 3:50:55 time: 1.3418 data_time: 0.0983 memory: 44222 loss_ce: 0.4148 loss: 0.4148 2022/09/15 18:59:49 - mmengine - INFO - Epoch(train) [2][1000/2304] lr: 1.0000e-03 eta: 3:46:22 time: 1.0529 data_time: 0.0065 memory: 44222 loss_ce: 0.4037 loss: 0.4037 2022/09/15 19:01:50 - mmengine - INFO - Epoch(train) [2][1100/2304] lr: 1.0000e-03 eta: 3:41:53 time: 1.0452 data_time: 0.0120 memory: 44222 loss_ce: 0.3938 loss: 0.3938 2022/09/15 19:03:53 - mmengine - INFO - Epoch(train) [2][1200/2304] lr: 1.0000e-03 eta: 3:37:35 time: 0.9337 data_time: 0.0180 memory: 44222 loss_ce: 0.3870 loss: 0.3870 2022/09/15 19:06:02 - mmengine - INFO - Epoch(train) [2][1300/2304] lr: 1.0000e-03 eta: 3:33:38 time: 1.4329 data_time: 0.3191 memory: 44222 loss_ce: 0.3977 loss: 0.3977 2022/09/15 19:08:07 - mmengine - INFO - Epoch(train) [2][1400/2304] lr: 1.0000e-03 eta: 3:29:38 time: 1.5024 data_time: 0.3467 memory: 44222 loss_ce: 0.4172 loss: 0.4172 2022/09/15 19:10:12 - mmengine - INFO - Epoch(train) [2][1500/2304] lr: 1.0000e-03 eta: 3:25:43 time: 1.3365 data_time: 0.0965 memory: 44222 loss_ce: 0.4045 loss: 0.4045 2022/09/15 19:12:16 - mmengine - INFO - Epoch(train) [2][1600/2304] lr: 1.0000e-03 eta: 3:21:52 time: 1.0635 data_time: 0.0065 memory: 44222 loss_ce: 0.3774 loss: 0.3774 2022/09/15 19:14:14 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 19:14:18 - mmengine - INFO - Epoch(train) [2][1700/2304] lr: 1.0000e-03 eta: 3:18:04 time: 1.0480 data_time: 0.0059 memory: 44222 loss_ce: 0.3803 loss: 0.3803 2022/09/15 19:16:21 - mmengine - INFO - Epoch(train) [2][1800/2304] lr: 1.0000e-03 eta: 3:14:22 time: 0.9589 data_time: 0.0202 memory: 44222 loss_ce: 0.3773 loss: 0.3773 2022/09/15 19:18:30 - mmengine - INFO - Epoch(train) [2][1900/2304] lr: 1.0000e-03 eta: 3:10:56 time: 1.4383 data_time: 0.3376 memory: 44222 loss_ce: 0.3752 loss: 0.3752 2022/09/15 19:20:33 - mmengine - INFO - Epoch(train) [2][2000/2304] lr: 1.0000e-03 eta: 3:07:24 time: 1.4707 data_time: 0.3205 memory: 44222 loss_ce: 0.3939 loss: 0.3939 2022/09/15 19:22:37 - mmengine - INFO - Epoch(train) [2][2100/2304] lr: 1.0000e-03 eta: 3:03:57 time: 1.3302 data_time: 0.1152 memory: 44222 loss_ce: 0.3725 loss: 0.3725 2022/09/15 19:24:44 - mmengine - INFO - Epoch(train) [2][2200/2304] lr: 1.0000e-03 eta: 3:00:37 time: 1.0407 data_time: 0.0056 memory: 44222 loss_ce: 0.3713 loss: 0.3713 2022/09/15 19:26:42 - mmengine - INFO - Epoch(train) [2][2300/2304] lr: 1.0000e-03 eta: 2:57:08 time: 0.9107 data_time: 0.0062 memory: 44222 loss_ce: 0.3577 loss: 0.3577 2022/09/15 19:26:45 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 19:26:45 - mmengine - INFO - Saving checkpoint at 2 epochs 2022/09/15 19:27:18 - mmengine - INFO - Epoch(val) [2][100/959] eta: 0:01:01 time: 0.0718 data_time: 0.0015 memory: 44222 2022/09/15 19:27:26 - mmengine - INFO - Epoch(val) [2][200/959] eta: 0:00:52 time: 0.0693 data_time: 0.0023 memory: 1130 2022/09/15 19:27:33 - mmengine - INFO - Epoch(val) [2][300/959] eta: 0:00:46 time: 0.0702 data_time: 0.0016 memory: 1130 2022/09/15 19:27:40 - mmengine - INFO - Epoch(val) [2][400/959] eta: 0:00:40 time: 0.0729 data_time: 0.0016 memory: 1130 2022/09/15 19:27:47 - mmengine - INFO - Epoch(val) [2][500/959] eta: 0:00:33 time: 0.0721 data_time: 0.0035 memory: 1130 2022/09/15 19:27:54 - mmengine - INFO - Epoch(val) [2][600/959] eta: 0:00:24 time: 0.0695 data_time: 0.0020 memory: 1130 2022/09/15 19:28:01 - mmengine - INFO - Epoch(val) [2][700/959] eta: 0:00:17 time: 0.0686 data_time: 0.0011 memory: 1130 2022/09/15 19:28:09 - mmengine - INFO - Epoch(val) [2][800/959] eta: 0:00:10 time: 0.0687 data_time: 0.0010 memory: 1130 2022/09/15 19:28:16 - mmengine - INFO - Epoch(val) [2][900/959] eta: 0:00:03 time: 0.0669 data_time: 0.0009 memory: 1130 2022/09/15 19:28:20 - mmengine - INFO - Epoch(val) [2][959/959] CUTE80/recog/word_acc_ignore_case_symbol: 0.8715 IIIT5K/recog/word_acc_ignore_case_symbol: 0.9383 SVT/recog/word_acc_ignore_case_symbol: 0.8794 SVTP/recog/word_acc_ignore_case_symbol: 0.7938 IC13/recog/word_acc_ignore_case_symbol: 0.9291 IC15/recog/word_acc_ignore_case_symbol: 0.7472 2022/09/15 19:30:34 - mmengine - INFO - Epoch(train) [3][100/2304] lr: 1.0000e-03 eta: 2:53:50 time: 1.4558 data_time: 0.2701 memory: 44222 loss_ce: 0.3561 loss: 0.3561 2022/09/15 19:32:41 - mmengine - INFO - Epoch(train) [3][200/2304] lr: 1.0000e-03 eta: 2:50:40 time: 1.7184 data_time: 0.3179 memory: 44222 loss_ce: 0.3594 loss: 0.3594 2022/09/15 19:34:46 - mmengine - INFO - Epoch(train) [3][300/2304] lr: 1.0000e-03 eta: 2:47:31 time: 1.4097 data_time: 0.1206 memory: 44222 loss_ce: 0.3739 loss: 0.3739 2022/09/15 19:36:43 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 19:36:50 - mmengine - INFO - Epoch(train) [3][400/2304] lr: 1.0000e-03 eta: 2:44:23 time: 1.0685 data_time: 0.0058 memory: 44222 loss_ce: 0.3443 loss: 0.3443 2022/09/15 19:38:55 - mmengine - INFO - Epoch(train) [3][500/2304] lr: 1.0000e-03 eta: 2:41:18 time: 0.9009 data_time: 0.0055 memory: 44222 loss_ce: 0.3485 loss: 0.3485 2022/09/15 19:40:59 - mmengine - INFO - Epoch(train) [3][600/2304] lr: 1.0000e-03 eta: 2:38:14 time: 0.8651 data_time: 0.0057 memory: 44222 loss_ce: 0.3464 loss: 0.3464 2022/09/15 19:43:12 - mmengine - INFO - Epoch(train) [3][700/2304] lr: 1.0000e-03 eta: 2:35:23 time: 1.4248 data_time: 0.2595 memory: 44222 loss_ce: 0.3511 loss: 0.3511 2022/09/15 19:45:19 - mmengine - INFO - Epoch(train) [3][800/2304] lr: 1.0000e-03 eta: 2:32:27 time: 1.7759 data_time: 0.3150 memory: 44222 loss_ce: 0.3417 loss: 0.3417 2022/09/15 19:47:24 - mmengine - INFO - Epoch(train) [3][900/2304] lr: 1.0000e-03 eta: 2:29:30 time: 1.4078 data_time: 0.1185 memory: 44222 loss_ce: 0.3479 loss: 0.3479 2022/09/15 19:49:29 - mmengine - INFO - Epoch(train) [3][1000/2304] lr: 1.0000e-03 eta: 2:26:36 time: 1.0838 data_time: 0.0067 memory: 44222 loss_ce: 0.3531 loss: 0.3531 2022/09/15 19:51:34 - mmengine - INFO - Epoch(train) [3][1100/2304] lr: 1.0000e-03 eta: 2:23:42 time: 0.9402 data_time: 0.0058 memory: 44222 loss_ce: 0.3792 loss: 0.3792 2022/09/15 19:53:38 - mmengine - INFO - Epoch(train) [3][1200/2304] lr: 1.0000e-03 eta: 2:20:50 time: 0.8733 data_time: 0.0065 memory: 44222 loss_ce: 0.3657 loss: 0.3657 2022/09/15 19:55:48 - mmengine - INFO - Epoch(train) [3][1300/2304] lr: 1.0000e-03 eta: 2:18:06 time: 1.4305 data_time: 0.2730 memory: 44222 loss_ce: 0.3503 loss: 0.3503 2022/09/15 19:57:41 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 19:57:58 - mmengine - INFO - Epoch(train) [3][1400/2304] lr: 1.0000e-03 eta: 2:15:21 time: 1.8258 data_time: 0.2962 memory: 44222 loss_ce: 0.3458 loss: 0.3458 2022/09/15 20:00:03 - mmengine - INFO - Epoch(train) [3][1500/2304] lr: 1.0000e-03 eta: 2:12:35 time: 1.4081 data_time: 0.1270 memory: 44222 loss_ce: 0.3443 loss: 0.3443 2022/09/15 20:02:10 - mmengine - INFO - Epoch(train) [3][1600/2304] lr: 1.0000e-03 eta: 2:09:51 time: 1.0954 data_time: 0.0064 memory: 44222 loss_ce: 0.3542 loss: 0.3542 2022/09/15 20:04:16 - mmengine - INFO - Epoch(train) [3][1700/2304] lr: 1.0000e-03 eta: 2:07:07 time: 0.9241 data_time: 0.0056 memory: 44222 loss_ce: 0.3548 loss: 0.3548 2022/09/15 20:06:21 - mmengine - INFO - Epoch(train) [3][1800/2304] lr: 1.0000e-03 eta: 2:04:23 time: 0.8485 data_time: 0.0059 memory: 44222 loss_ce: 0.3446 loss: 0.3446 2022/09/15 20:08:32 - mmengine - INFO - Epoch(train) [3][1900/2304] lr: 1.0000e-03 eta: 2:01:46 time: 1.4546 data_time: 0.2654 memory: 44222 loss_ce: 0.3340 loss: 0.3340 2022/09/15 20:10:41 - mmengine - INFO - Epoch(train) [3][2000/2304] lr: 1.0000e-03 eta: 1:59:07 time: 1.8265 data_time: 0.3281 memory: 44222 loss_ce: 0.3422 loss: 0.3422 2022/09/15 20:12:47 - mmengine - INFO - Epoch(train) [3][2100/2304] lr: 1.0000e-03 eta: 1:56:28 time: 1.4121 data_time: 0.1128 memory: 44222 loss_ce: 0.3522 loss: 0.3522 2022/09/15 20:14:52 - mmengine - INFO - Epoch(train) [3][2200/2304] lr: 1.0000e-03 eta: 1:53:49 time: 1.0504 data_time: 0.0067 memory: 44222 loss_ce: 0.3571 loss: 0.3571 2022/09/15 20:16:56 - mmengine - INFO - Epoch(train) [3][2300/2304] lr: 1.0000e-03 eta: 1:51:10 time: 0.9038 data_time: 0.0068 memory: 44222 loss_ce: 0.3462 loss: 0.3462 2022/09/15 20:16:59 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 20:16:59 - mmengine - INFO - Saving checkpoint at 3 epochs 2022/09/15 20:17:27 - mmengine - INFO - Epoch(val) [3][100/959] eta: 0:01:02 time: 0.0725 data_time: 0.0022 memory: 44222 2022/09/15 20:17:35 - mmengine - INFO - Epoch(val) [3][200/959] eta: 0:00:56 time: 0.0751 data_time: 0.0031 memory: 1130 2022/09/15 20:17:42 - mmengine - INFO - Epoch(val) [3][300/959] eta: 0:00:46 time: 0.0701 data_time: 0.0009 memory: 1130 2022/09/15 20:17:49 - mmengine - INFO - Epoch(val) [3][400/959] eta: 0:00:37 time: 0.0666 data_time: 0.0021 memory: 1130 2022/09/15 20:17:56 - mmengine - INFO - Epoch(val) [3][500/959] eta: 0:00:32 time: 0.0715 data_time: 0.0020 memory: 1130 2022/09/15 20:18:04 - mmengine - INFO - Epoch(val) [3][600/959] eta: 0:00:25 time: 0.0701 data_time: 0.0010 memory: 1130 2022/09/15 20:18:10 - mmengine - INFO - Epoch(val) [3][700/959] eta: 0:00:17 time: 0.0678 data_time: 0.0033 memory: 1130 2022/09/15 20:18:18 - mmengine - INFO - Epoch(val) [3][800/959] eta: 0:00:11 time: 0.0721 data_time: 0.0023 memory: 1130 2022/09/15 20:18:24 - mmengine - INFO - Epoch(val) [3][900/959] eta: 0:00:03 time: 0.0658 data_time: 0.0006 memory: 1130 2022/09/15 20:18:28 - mmengine - INFO - Epoch(val) [3][959/959] CUTE80/recog/word_acc_ignore_case_symbol: 0.8785 IIIT5K/recog/word_acc_ignore_case_symbol: 0.9453 SVT/recog/word_acc_ignore_case_symbol: 0.8733 SVTP/recog/word_acc_ignore_case_symbol: 0.7938 IC13/recog/word_acc_ignore_case_symbol: 0.9222 IC15/recog/word_acc_ignore_case_symbol: 0.7352 2022/09/15 20:20:25 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 20:20:40 - mmengine - INFO - Epoch(train) [4][100/2304] lr: 1.0000e-04 eta: 1:48:27 time: 1.3783 data_time: 0.1985 memory: 44222 loss_ce: 0.3102 loss: 0.3102 2022/09/15 20:22:45 - mmengine - INFO - Epoch(train) [4][200/2304] lr: 1.0000e-04 eta: 1:45:51 time: 1.6075 data_time: 0.2593 memory: 44222 loss_ce: 0.3153 loss: 0.3153 2022/09/15 20:24:48 - mmengine - INFO - Epoch(train) [4][300/2304] lr: 1.0000e-04 eta: 1:43:14 time: 1.4008 data_time: 0.1956 memory: 44222 loss_ce: 0.2994 loss: 0.2994 2022/09/15 20:26:51 - mmengine - INFO - Epoch(train) [4][400/2304] lr: 1.0000e-04 eta: 1:40:39 time: 1.0621 data_time: 0.1173 memory: 44222 loss_ce: 0.3036 loss: 0.3036 2022/09/15 20:28:54 - mmengine - INFO - Epoch(train) [4][500/2304] lr: 1.0000e-04 eta: 1:38:04 time: 0.9892 data_time: 0.0505 memory: 44222 loss_ce: 0.3245 loss: 0.3245 2022/09/15 20:30:56 - mmengine - INFO - Epoch(train) [4][600/2304] lr: 1.0000e-04 eta: 1:35:29 time: 0.8926 data_time: 0.0186 memory: 44222 loss_ce: 0.2988 loss: 0.2988 2022/09/15 20:33:04 - mmengine - INFO - Epoch(train) [4][700/2304] lr: 1.0000e-04 eta: 1:32:58 time: 1.3811 data_time: 0.1849 memory: 44222 loss_ce: 0.3044 loss: 0.3044 2022/09/15 20:35:11 - mmengine - INFO - Epoch(train) [4][800/2304] lr: 1.0000e-04 eta: 1:30:28 time: 1.6325 data_time: 0.2193 memory: 44222 loss_ce: 0.3002 loss: 0.3002 2022/09/15 20:37:16 - mmengine - INFO - Epoch(train) [4][900/2304] lr: 1.0000e-04 eta: 1:27:57 time: 1.4185 data_time: 0.2106 memory: 44222 loss_ce: 0.3120 loss: 0.3120 2022/09/15 20:39:21 - mmengine - INFO - Epoch(train) [4][1000/2304] lr: 1.0000e-04 eta: 1:25:27 time: 1.0741 data_time: 0.0928 memory: 44222 loss_ce: 0.3076 loss: 0.3076 2022/09/15 20:41:14 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 20:41:26 - mmengine - INFO - Epoch(train) [4][1100/2304] lr: 1.0000e-04 eta: 1:22:57 time: 0.9921 data_time: 0.0691 memory: 44222 loss_ce: 0.3092 loss: 0.3092 2022/09/15 20:43:30 - mmengine - INFO - Epoch(train) [4][1200/2304] lr: 1.0000e-04 eta: 1:20:28 time: 0.8913 data_time: 0.0332 memory: 44222 loss_ce: 0.3048 loss: 0.3048 2022/09/15 20:45:39 - mmengine - INFO - Epoch(train) [4][1300/2304] lr: 1.0000e-04 eta: 1:18:01 time: 1.4362 data_time: 0.2003 memory: 44222 loss_ce: 0.3148 loss: 0.3148 2022/09/15 20:47:46 - mmengine - INFO - Epoch(train) [4][1400/2304] lr: 1.0000e-04 eta: 1:15:34 time: 1.6249 data_time: 0.2376 memory: 44222 loss_ce: 0.3162 loss: 0.3162 2022/09/15 20:49:52 - mmengine - INFO - Epoch(train) [4][1500/2304] lr: 1.0000e-04 eta: 1:13:07 time: 1.4339 data_time: 0.1996 memory: 44222 loss_ce: 0.2990 loss: 0.2990 2022/09/15 20:52:00 - mmengine - INFO - Epoch(train) [4][1600/2304] lr: 1.0000e-04 eta: 1:10:41 time: 1.1013 data_time: 0.0950 memory: 44222 loss_ce: 0.3082 loss: 0.3082 2022/09/15 20:54:05 - mmengine - INFO - Epoch(train) [4][1700/2304] lr: 1.0000e-04 eta: 1:08:15 time: 0.9470 data_time: 0.0329 memory: 44222 loss_ce: 0.3060 loss: 0.3060 2022/09/15 20:56:09 - mmengine - INFO - Epoch(train) [4][1800/2304] lr: 1.0000e-04 eta: 1:05:48 time: 0.8722 data_time: 0.0222 memory: 44222 loss_ce: 0.3044 loss: 0.3044 2022/09/15 20:58:19 - mmengine - INFO - Epoch(train) [4][1900/2304] lr: 1.0000e-04 eta: 1:03:24 time: 1.4251 data_time: 0.2075 memory: 44222 loss_ce: 0.2952 loss: 0.2952 2022/09/15 21:00:26 - mmengine - INFO - Epoch(train) [4][2000/2304] lr: 1.0000e-04 eta: 1:01:00 time: 1.6760 data_time: 0.2578 memory: 44222 loss_ce: 0.2934 loss: 0.2934 2022/09/15 21:02:13 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 21:02:33 - mmengine - INFO - Epoch(train) [4][2100/2304] lr: 1.0000e-04 eta: 0:58:36 time: 1.4361 data_time: 0.2413 memory: 44222 loss_ce: 0.3065 loss: 0.3065 2022/09/15 21:04:37 - mmengine - INFO - Epoch(train) [4][2200/2304] lr: 1.0000e-04 eta: 0:56:11 time: 1.1021 data_time: 0.1004 memory: 44222 loss_ce: 0.3015 loss: 0.3015 2022/09/15 21:06:38 - mmengine - INFO - Epoch(train) [4][2300/2304] lr: 1.0000e-04 eta: 0:53:46 time: 0.9167 data_time: 0.0285 memory: 44222 loss_ce: 0.3060 loss: 0.3060 2022/09/15 21:06:41 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 21:06:41 - mmengine - INFO - Saving checkpoint at 4 epochs 2022/09/15 21:07:10 - mmengine - INFO - Epoch(val) [4][100/959] eta: 0:00:59 time: 0.0692 data_time: 0.0029 memory: 44222 2022/09/15 21:07:17 - mmengine - INFO - Epoch(val) [4][200/959] eta: 0:00:55 time: 0.0737 data_time: 0.0016 memory: 1130 2022/09/15 21:07:24 - mmengine - INFO - Epoch(val) [4][300/959] eta: 0:00:45 time: 0.0695 data_time: 0.0013 memory: 1130 2022/09/15 21:07:31 - mmengine - INFO - Epoch(val) [4][400/959] eta: 0:00:40 time: 0.0730 data_time: 0.0023 memory: 1130 2022/09/15 21:07:38 - mmengine - INFO - Epoch(val) [4][500/959] eta: 0:00:31 time: 0.0694 data_time: 0.0014 memory: 1130 2022/09/15 21:07:46 - mmengine - INFO - Epoch(val) [4][600/959] eta: 0:00:24 time: 0.0693 data_time: 0.0031 memory: 1130 2022/09/15 21:07:53 - mmengine - INFO - Epoch(val) [4][700/959] eta: 0:00:17 time: 0.0689 data_time: 0.0011 memory: 1130 2022/09/15 21:08:00 - mmengine - INFO - Epoch(val) [4][800/959] eta: 0:00:11 time: 0.0700 data_time: 0.0013 memory: 1130 2022/09/15 21:08:06 - mmengine - INFO - Epoch(val) [4][900/959] eta: 0:00:03 time: 0.0664 data_time: 0.0019 memory: 1130 2022/09/15 21:08:10 - mmengine - INFO - Epoch(val) [4][959/959] CUTE80/recog/word_acc_ignore_case_symbol: 0.8958 IIIT5K/recog/word_acc_ignore_case_symbol: 0.9550 SVT/recog/word_acc_ignore_case_symbol: 0.8825 SVTP/recog/word_acc_ignore_case_symbol: 0.8233 IC13/recog/word_acc_ignore_case_symbol: 0.9369 IC15/recog/word_acc_ignore_case_symbol: 0.7641 2022/09/15 21:10:29 - mmengine - INFO - Epoch(train) [5][100/2304] lr: 1.0000e-05 eta: 0:51:19 time: 1.6232 data_time: 0.2689 memory: 44222 loss_ce: 0.3038 loss: 0.3038 2022/09/15 21:12:35 - mmengine - INFO - Epoch(train) [5][200/2304] lr: 1.0000e-05 eta: 0:48:57 time: 1.7067 data_time: 0.2649 memory: 44222 loss_ce: 0.2934 loss: 0.2934 2022/09/15 21:14:39 - mmengine - INFO - Epoch(train) [5][300/2304] lr: 1.0000e-05 eta: 0:46:34 time: 1.4301 data_time: 0.2168 memory: 44222 loss_ce: 0.3017 loss: 0.3017 2022/09/15 21:16:42 - mmengine - INFO - Epoch(train) [5][400/2304] lr: 1.0000e-05 eta: 0:44:11 time: 0.9670 data_time: 0.0293 memory: 44222 loss_ce: 0.2985 loss: 0.2985 2022/09/15 21:18:45 - mmengine - INFO - Epoch(train) [5][500/2304] lr: 1.0000e-05 eta: 0:41:49 time: 0.8828 data_time: 0.0066 memory: 44222 loss_ce: 0.2874 loss: 0.2874 2022/09/15 21:20:48 - mmengine - INFO - Epoch(train) [5][600/2304] lr: 1.0000e-05 eta: 0:39:27 time: 0.8355 data_time: 0.0060 memory: 44222 loss_ce: 0.2964 loss: 0.2964 2022/09/15 21:22:58 - mmengine - INFO - Epoch(train) [5][700/2304] lr: 1.0000e-05 eta: 0:37:07 time: 1.5185 data_time: 0.2848 memory: 44222 loss_ce: 0.2881 loss: 0.2881 2022/09/15 21:24:41 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 21:25:04 - mmengine - INFO - Epoch(train) [5][800/2304] lr: 1.0000e-05 eta: 0:34:46 time: 1.7960 data_time: 0.2940 memory: 44222 loss_ce: 0.2981 loss: 0.2981 2022/09/15 21:27:08 - mmengine - INFO - Epoch(train) [5][900/2304] lr: 1.0000e-05 eta: 0:32:25 time: 1.3812 data_time: 0.1936 memory: 44222 loss_ce: 0.2989 loss: 0.2989 2022/09/15 21:29:12 - mmengine - INFO - Epoch(train) [5][1000/2304] lr: 1.0000e-05 eta: 0:30:05 time: 0.9796 data_time: 0.0268 memory: 44222 loss_ce: 0.2936 loss: 0.2936 2022/09/15 21:31:15 - mmengine - INFO - Epoch(train) [5][1100/2304] lr: 1.0000e-05 eta: 0:27:44 time: 0.8753 data_time: 0.0066 memory: 44222 loss_ce: 0.3069 loss: 0.3069 2022/09/15 21:33:19 - mmengine - INFO - Epoch(train) [5][1200/2304] lr: 1.0000e-05 eta: 0:25:24 time: 0.8444 data_time: 0.0065 memory: 44222 loss_ce: 0.2939 loss: 0.2939 2022/09/15 21:35:28 - mmengine - INFO - Epoch(train) [5][1300/2304] lr: 1.0000e-05 eta: 0:23:06 time: 1.4760 data_time: 0.2537 memory: 44222 loss_ce: 0.2950 loss: 0.2950 2022/09/15 21:37:34 - mmengine - INFO - Epoch(train) [5][1400/2304] lr: 1.0000e-05 eta: 0:20:46 time: 1.7002 data_time: 0.3021 memory: 44222 loss_ce: 0.2862 loss: 0.2862 2022/09/15 21:39:40 - mmengine - INFO - Epoch(train) [5][1500/2304] lr: 1.0000e-05 eta: 0:18:28 time: 1.4669 data_time: 0.2081 memory: 44222 loss_ce: 0.2888 loss: 0.2888 2022/09/15 21:41:45 - mmengine - INFO - Epoch(train) [5][1600/2304] lr: 1.0000e-05 eta: 0:16:09 time: 1.0049 data_time: 0.0515 memory: 44222 loss_ce: 0.2926 loss: 0.2926 2022/09/15 21:43:51 - mmengine - INFO - Epoch(train) [5][1700/2304] lr: 1.0000e-05 eta: 0:13:51 time: 0.8519 data_time: 0.0063 memory: 44222 loss_ce: 0.2909 loss: 0.2909 2022/09/15 21:45:42 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 21:45:56 - mmengine - INFO - Epoch(train) [5][1800/2304] lr: 1.0000e-05 eta: 0:11:32 time: 0.8522 data_time: 0.0060 memory: 44222 loss_ce: 0.2990 loss: 0.2990 2022/09/15 21:48:06 - mmengine - INFO - Epoch(train) [5][1900/2304] lr: 1.0000e-05 eta: 0:09:15 time: 1.5017 data_time: 0.2855 memory: 44222 loss_ce: 0.3046 loss: 0.3046 2022/09/15 21:50:14 - mmengine - INFO - Epoch(train) [5][2000/2304] lr: 1.0000e-05 eta: 0:06:57 time: 1.7874 data_time: 0.3002 memory: 44222 loss_ce: 0.2915 loss: 0.2915 2022/09/15 21:52:20 - mmengine - INFO - Epoch(train) [5][2100/2304] lr: 1.0000e-05 eta: 0:04:39 time: 1.4280 data_time: 0.1991 memory: 44222 loss_ce: 0.2998 loss: 0.2998 2022/09/15 21:54:23 - mmengine - INFO - Epoch(train) [5][2200/2304] lr: 1.0000e-05 eta: 0:02:22 time: 0.9380 data_time: 0.0491 memory: 44222 loss_ce: 0.3065 loss: 0.3065 2022/09/15 21:56:22 - mmengine - INFO - Epoch(train) [5][2300/2304] lr: 1.0000e-05 eta: 0:00:05 time: 0.8102 data_time: 0.0082 memory: 44222 loss_ce: 0.3043 loss: 0.3043 2022/09/15 21:56:26 - mmengine - INFO - Exp name: sar_resnet31_parallel-decoder_5e_st-sub_mj-sub_sa_real_20220915_171910 2022/09/15 21:56:26 - mmengine - INFO - Saving checkpoint at 5 epochs 2022/09/15 21:56:54 - mmengine - INFO - Epoch(val) [5][100/959] eta: 0:00:57 time: 0.0674 data_time: 0.0012 memory: 44222 2022/09/15 21:57:01 - mmengine - INFO - Epoch(val) [5][200/959] eta: 0:00:54 time: 0.0716 data_time: 0.0030 memory: 1130 2022/09/15 21:57:08 - mmengine - INFO - Epoch(val) [5][300/959] eta: 0:00:49 time: 0.0744 data_time: 0.0033 memory: 1130 2022/09/15 21:57:16 - mmengine - INFO - Epoch(val) [5][400/959] eta: 0:00:38 time: 0.0687 data_time: 0.0017 memory: 1130 2022/09/15 21:57:23 - mmengine - INFO - Epoch(val) [5][500/959] eta: 0:00:32 time: 0.0701 data_time: 0.0011 memory: 1130 2022/09/15 21:57:30 - mmengine - INFO - Epoch(val) [5][600/959] eta: 0:00:23 time: 0.0663 data_time: 0.0013 memory: 1130 2022/09/15 21:57:37 - mmengine - INFO - Epoch(val) [5][700/959] eta: 0:00:18 time: 0.0713 data_time: 0.0026 memory: 1130 2022/09/15 21:57:44 - mmengine - INFO - Epoch(val) [5][800/959] eta: 0:00:12 time: 0.0795 data_time: 0.0022 memory: 1130 2022/09/15 21:57:51 - mmengine - INFO - Epoch(val) [5][900/959] eta: 0:00:04 time: 0.0700 data_time: 0.0014 memory: 1130 2022/09/15 21:57:55 - mmengine - INFO - Epoch(val) [5][959/959] CUTE80/recog/word_acc_ignore_case_symbol: 0.9028 IIIT5K/recog/word_acc_ignore_case_symbol: 0.9533 SVT/recog/word_acc_ignore_case_symbol: 0.8841 SVTP/recog/word_acc_ignore_case_symbol: 0.8326 IC13/recog/word_acc_ignore_case_symbol: 0.9369 IC15/recog/word_acc_ignore_case_symbol: 0.7602