2023/03/06 23:23:00 - mmengine - INFO - ------------------------------------------------------------ System environment: sys.platform: linux Python: 3.9.13 (main, Aug 25 2022, 23:26:10) [GCC 11.2.0] CUDA available: True numpy_random_seed: 992735325 GPU 0,1,2,3,4,5,6,7: NVIDIA A100-SXM4-80GB CUDA_HOME: /mnt/petrelfs/share/cuda-11.3 NVCC: Cuda compilation tools, release 11.3, V11.3.109 GCC: gcc (GCC) 5.4.0 PyTorch: 1.11.0 PyTorch compiling details: PyTorch built with: - GCC 7.3 - C++ Version: 201402 - Intel(R) oneAPI Math Kernel Library Version 2021.4-Product Build 20210904 for Intel(R) 64 architecture applications - Intel(R) MKL-DNN v2.5.2 (Git Hash a9302535553c73243c632ad3c4c80beec3d19a1e) - OpenMP 201511 (a.k.a. OpenMP 4.5) - LAPACK is enabled (usually provided by MKL) - NNPACK is enabled - CPU capability usage: AVX2 - CUDA Runtime 11.3 - 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_61,code=sm_61;-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;-gencode;arch=compute_37,code=compute_37 - CuDNN 8.2 - Magma 2.5.2 - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=11.3, CUDNN_VERSION=8.2.0, 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 -DEDGE_PROFILER_USE_KINETO -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-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=1.11.0, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=OFF, USE_MPI=OFF, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, TorchVision: 0.12.0 OpenCV: 4.6.0 MMEngine: 0.6.0 Runtime environment: cudnn_benchmark: False mp_cfg: {'mp_start_method': 'fork', 'opencv_num_threads': 0} dist_cfg: {'backend': 'nccl'} seed: None diff_rank_seed: False deterministic: False Distributed launcher: slurm Distributed training: True GPU number: 8 ------------------------------------------------------------ 2023/03/06 23:23:00 - mmengine - INFO - Config: default_scope = 'mmaction' default_hooks = dict( runtime_info=dict(type='RuntimeInfoHook', _scope_='mmaction'), timer=dict(type='IterTimerHook', _scope_='mmaction'), logger=dict( type='LoggerHook', interval=100, ignore_last=False, _scope_='mmaction'), param_scheduler=dict(type='ParamSchedulerHook', _scope_='mmaction'), checkpoint=dict( type='CheckpointHook', interval=1, save_best='auto', _scope_='mmaction'), sampler_seed=dict(type='DistSamplerSeedHook', _scope_='mmaction'), sync_buffers=dict(type='SyncBuffersHook', _scope_='mmaction')) env_cfg = dict( cudnn_benchmark=False, mp_cfg=dict(mp_start_method='fork', opencv_num_threads=0), dist_cfg=dict(backend='nccl')) log_processor = dict( type='LogProcessor', window_size=20, by_epoch=True, _scope_='mmaction') vis_backends = [dict(type='LocalVisBackend', _scope_='mmaction')] visualizer = dict( type='ActionVisualizer', vis_backends=[dict(type='LocalVisBackend')], _scope_='mmaction') log_level = 'INFO' load_from = None resume = False custom_imports = dict(imports='models') model = dict( type='RecognizerGCN', backbone=dict( type='CTRGCN', graph_cfg=dict(layout='nturgb+d', mode='spatial')), cls_head=dict(type='GCNHead', num_classes=60, in_channels=256)) dataset_type = 'PoseDataset' ann_file = 'data/skeleton/ntu60_3d.pkl' train_pipeline = [ dict(type='PreNormalize3D'), dict(type='GenSkeFeat', dataset='nturgb+d', feats=['j']), dict(type='UniformSampleFrames', clip_len=100), dict(type='PoseDecode'), dict(type='FormatGCNInput', num_person=2), dict(type='PackActionInputs') ] val_pipeline = [ dict(type='PreNormalize3D'), dict(type='GenSkeFeat', dataset='nturgb+d', feats=['j']), dict( type='UniformSampleFrames', clip_len=100, num_clips=1, test_mode=True), dict(type='PoseDecode'), dict(type='FormatGCNInput', num_person=2), dict(type='PackActionInputs') ] test_pipeline = [ dict(type='PreNormalize3D'), dict(type='GenSkeFeat', dataset='nturgb+d', feats=['j']), dict( type='UniformSampleFrames', clip_len=100, num_clips=10, test_mode=True), dict(type='PoseDecode'), dict(type='FormatGCNInput', num_person=2), dict(type='PackActionInputs') ] train_dataloader = dict( batch_size=16, num_workers=2, persistent_workers=True, sampler=dict(type='DefaultSampler', shuffle=True), dataset=dict( type='RepeatDataset', times=5, dataset=dict( type='PoseDataset', ann_file='data/skeleton/ntu60_3d.pkl', pipeline=[ dict(type='PreNormalize3D'), dict(type='GenSkeFeat', dataset='nturgb+d', feats=['j']), dict(type='UniformSampleFrames', clip_len=100), dict(type='PoseDecode'), dict(type='FormatGCNInput', num_person=2), dict(type='PackActionInputs') ], split='xsub_train'))) val_dataloader = dict( batch_size=16, num_workers=2, persistent_workers=True, sampler=dict(type='DefaultSampler', shuffle=False), dataset=dict( type='PoseDataset', ann_file='data/skeleton/ntu60_3d.pkl', pipeline=[ dict(type='PreNormalize3D'), dict(type='GenSkeFeat', dataset='nturgb+d', feats=['j']), dict( type='UniformSampleFrames', clip_len=100, num_clips=1, test_mode=True), dict(type='PoseDecode'), dict(type='FormatGCNInput', num_person=2), dict(type='PackActionInputs') ], split='xsub_val', test_mode=True)) test_dataloader = dict( batch_size=1, num_workers=2, persistent_workers=True, sampler=dict(type='DefaultSampler', shuffle=False), dataset=dict( type='PoseDataset', ann_file='data/skeleton/ntu60_3d.pkl', pipeline=[ dict(type='PreNormalize3D'), dict(type='GenSkeFeat', dataset='nturgb+d', feats=['j']), dict( type='UniformSampleFrames', clip_len=100, num_clips=10, test_mode=True), dict(type='PoseDecode'), dict(type='FormatGCNInput', num_person=2), dict(type='PackActionInputs') ], split='xsub_val', test_mode=True)) val_evaluator = [dict(type='AccMetric')] test_evaluator = [dict(type='AccMetric')] train_cfg = dict( type='EpochBasedTrainLoop', max_epochs=16, val_begin=1, val_interval=1) val_cfg = dict(type='ValLoop') test_cfg = dict(type='TestLoop') param_scheduler = [ dict( type='CosineAnnealingLR', eta_min=0, T_max=16, by_epoch=True, convert_to_iter_based=True) ] optim_wrapper = dict( optimizer=dict( type='SGD', lr=0.1, momentum=0.9, weight_decay=0.0005, nesterov=True)) auto_scale_lr = dict(enable=False, base_batch_size=128) launcher = 'slurm' work_dir = './work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d' randomness = dict(seed=None, diff_rank_seed=False, deterministic=False) 2023/03/06 23:23:01 - mmengine - INFO - Hooks will be executed in the following order: before_run: (VERY_HIGH ) RuntimeInfoHook (BELOW_NORMAL) LoggerHook -------------------- before_train: (VERY_HIGH ) RuntimeInfoHook (NORMAL ) IterTimerHook (VERY_LOW ) CheckpointHook -------------------- before_train_epoch: (VERY_HIGH ) RuntimeInfoHook (NORMAL ) IterTimerHook (NORMAL ) DistSamplerSeedHook -------------------- before_train_iter: (VERY_HIGH ) RuntimeInfoHook (NORMAL ) IterTimerHook -------------------- after_train_iter: (VERY_HIGH ) RuntimeInfoHook (NORMAL ) IterTimerHook (BELOW_NORMAL) LoggerHook (LOW ) ParamSchedulerHook (VERY_LOW ) CheckpointHook -------------------- after_train_epoch: (NORMAL ) IterTimerHook (NORMAL ) SyncBuffersHook (LOW ) ParamSchedulerHook (VERY_LOW ) CheckpointHook -------------------- before_val_epoch: (NORMAL ) IterTimerHook -------------------- before_val_iter: (NORMAL ) IterTimerHook -------------------- after_val_iter: (NORMAL ) IterTimerHook (BELOW_NORMAL) LoggerHook -------------------- after_val_epoch: (VERY_HIGH ) RuntimeInfoHook (NORMAL ) IterTimerHook (BELOW_NORMAL) LoggerHook (LOW ) ParamSchedulerHook (VERY_LOW ) CheckpointHook -------------------- before_test_epoch: (NORMAL ) IterTimerHook -------------------- before_test_iter: (NORMAL ) IterTimerHook -------------------- after_test_iter: (NORMAL ) IterTimerHook (BELOW_NORMAL) LoggerHook -------------------- after_test_epoch: (VERY_HIGH ) RuntimeInfoHook (NORMAL ) IterTimerHook (BELOW_NORMAL) LoggerHook -------------------- after_run: (BELOW_NORMAL) LoggerHook -------------------- Name of parameter - Initialization information backbone.data_bn.weight - torch.Size([150]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.data_bn.bias - torch.Size([150]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.0.conv1.weight - torch.Size([8, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.0.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.0.conv2.weight - torch.Size([8, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.0.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.0.conv3.weight - torch.Size([64, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.0.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.0.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.0.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.1.conv1.weight - torch.Size([8, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.1.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.1.conv2.weight - torch.Size([8, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.1.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.1.conv3.weight - torch.Size([64, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.1.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.1.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.1.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.2.conv1.weight - torch.Size([8, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.2.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.2.conv2.weight - torch.Size([8, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.2.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.2.conv3.weight - torch.Size([64, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.2.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.convs.2.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.convs.2.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.down.0.weight - torch.Size([64, 3, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.down.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.gcn1.down.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.down.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.gcn1.bn.weight - torch.Size([64]): ConstantInit: val=1e-06, bias=0 backbone.net.0.gcn1.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.0.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.0.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.0.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.0.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.0.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.0.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.0.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.0.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.1.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.1.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.1.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.1.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.1.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.1.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.1.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.1.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.2.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.2.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.2.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.2.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.2.4.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.2.4.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.3.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.3.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.0.tcn1.branches.3.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.0.tcn1.branches.3.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.0.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.0.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.0.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.0.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.0.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.0.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.0.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.0.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.1.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.1.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.1.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.1.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.1.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.1.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.1.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.1.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.2.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.2.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.2.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.2.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.2.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.2.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.convs.2.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.gcn1.convs.2.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.gcn1.bn.weight - torch.Size([64]): ConstantInit: val=1e-06, bias=0 backbone.net.1.gcn1.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.0.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.0.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.0.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.0.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.0.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.0.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.0.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.0.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.1.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.1.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.1.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.1.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.1.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.1.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.1.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.1.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.2.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.2.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.2.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.2.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.2.4.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.2.4.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.3.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.3.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.1.tcn1.branches.3.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.1.tcn1.branches.3.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.0.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.0.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.0.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.0.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.0.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.0.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.0.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.0.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.1.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.1.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.1.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.1.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.1.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.1.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.1.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.1.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.2.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.2.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.2.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.2.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.2.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.2.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.convs.2.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.gcn1.convs.2.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.gcn1.bn.weight - torch.Size([64]): ConstantInit: val=1e-06, bias=0 backbone.net.2.gcn1.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.0.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.0.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.0.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.0.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.0.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.0.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.0.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.0.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.1.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.1.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.1.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.1.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.1.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.1.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.1.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.1.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.2.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.2.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.2.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.2.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.2.4.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.2.4.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.3.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.3.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.2.tcn1.branches.3.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.2.tcn1.branches.3.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.0.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.0.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.0.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.0.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.0.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.0.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.0.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.0.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.1.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.1.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.1.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.1.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.1.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.1.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.1.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.1.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.2.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.2.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.2.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.2.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.2.conv3.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.2.conv3.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.convs.2.conv4.weight - torch.Size([64, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.gcn1.convs.2.conv4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.gcn1.bn.weight - torch.Size([64]): ConstantInit: val=1e-06, bias=0 backbone.net.3.gcn1.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.0.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.0.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.0.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.0.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.0.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.0.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.0.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.0.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.1.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.1.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.1.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.1.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.1.3.conv.weight - torch.Size([16, 16, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.1.3.conv.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.1.3.bn.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.1.3.bn.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.2.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.2.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.2.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.2.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.2.4.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.2.4.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.3.0.weight - torch.Size([16, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.3.0.bias - torch.Size([16]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.3.tcn1.branches.3.1.weight - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.3.tcn1.branches.3.1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.0.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.0.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.0.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.0.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.0.conv3.weight - torch.Size([128, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.0.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.0.conv4.weight - torch.Size([128, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.0.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.1.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.1.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.1.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.1.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.1.conv3.weight - torch.Size([128, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.1.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.1.conv4.weight - torch.Size([128, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.1.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.2.conv1.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.2.conv1.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.2.conv2.weight - torch.Size([8, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.2.conv2.bias - torch.Size([8]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.2.conv3.weight - torch.Size([128, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.2.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.convs.2.conv4.weight - torch.Size([128, 8, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.convs.2.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.down.0.weight - torch.Size([128, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.down.0.bias - torch.Size([128]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.gcn1.down.1.weight - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.down.1.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.gcn1.bn.weight - torch.Size([128]): ConstantInit: val=1e-06, bias=0 backbone.net.4.gcn1.bn.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.0.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.0.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.0.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.0.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.0.3.conv.weight - torch.Size([32, 32, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.0.3.conv.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.0.3.bn.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.0.3.bn.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.1.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.1.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.1.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.1.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.1.3.conv.weight - torch.Size([32, 32, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.1.3.conv.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.1.3.bn.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.1.3.bn.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.2.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.2.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.2.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.2.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.2.4.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.2.4.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.3.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.3.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.tcn1.branches.3.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.tcn1.branches.3.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.residual.conv.weight - torch.Size([128, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.residual.conv.bias - torch.Size([128]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.4.residual.bn.weight - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.4.residual.bn.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.0.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.0.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.0.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.0.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.0.conv3.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.0.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.0.conv4.weight - torch.Size([128, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.0.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.1.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.1.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.1.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.1.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.1.conv3.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.1.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.1.conv4.weight - torch.Size([128, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.1.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.2.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.2.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.2.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.2.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.2.conv3.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.2.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.convs.2.conv4.weight - torch.Size([128, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.gcn1.convs.2.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.gcn1.bn.weight - torch.Size([128]): ConstantInit: val=1e-06, bias=0 backbone.net.5.gcn1.bn.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.0.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.0.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.0.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.0.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.0.3.conv.weight - torch.Size([32, 32, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.0.3.conv.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.0.3.bn.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.0.3.bn.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.1.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.1.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.1.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.1.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.1.3.conv.weight - torch.Size([32, 32, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.1.3.conv.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.1.3.bn.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.1.3.bn.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.2.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.2.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.2.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.2.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.2.4.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.2.4.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.3.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.3.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.5.tcn1.branches.3.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.5.tcn1.branches.3.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.0.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.0.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.0.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.0.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.0.conv3.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.0.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.0.conv4.weight - torch.Size([128, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.0.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.1.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.1.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.1.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.1.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.1.conv3.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.1.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.1.conv4.weight - torch.Size([128, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.1.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.2.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.2.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.2.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.2.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.2.conv3.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.2.conv3.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.convs.2.conv4.weight - torch.Size([128, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.gcn1.convs.2.conv4.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.gcn1.bn.weight - torch.Size([128]): ConstantInit: val=1e-06, bias=0 backbone.net.6.gcn1.bn.bias - torch.Size([128]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.0.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.0.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.0.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.0.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.0.3.conv.weight - torch.Size([32, 32, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.0.3.conv.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.0.3.bn.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.0.3.bn.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.1.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.1.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.1.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.1.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.1.3.conv.weight - torch.Size([32, 32, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.1.3.conv.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.1.3.bn.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.1.3.bn.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.2.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.2.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.2.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.2.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.2.4.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.2.4.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.3.0.weight - torch.Size([32, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.3.0.bias - torch.Size([32]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.6.tcn1.branches.3.1.weight - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.6.tcn1.branches.3.1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.0.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.0.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.0.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.0.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.0.conv3.weight - torch.Size([256, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.0.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.0.conv4.weight - torch.Size([256, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.0.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.1.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.1.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.1.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.1.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.1.conv3.weight - torch.Size([256, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.1.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.1.conv4.weight - torch.Size([256, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.1.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.2.conv1.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.2.conv1.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.2.conv2.weight - torch.Size([16, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.2.conv2.bias - torch.Size([16]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.2.conv3.weight - torch.Size([256, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.2.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.convs.2.conv4.weight - torch.Size([256, 16, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.convs.2.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.down.0.weight - torch.Size([256, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.down.0.bias - torch.Size([256]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.gcn1.down.1.weight - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.down.1.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.gcn1.bn.weight - torch.Size([256]): ConstantInit: val=1e-06, bias=0 backbone.net.7.gcn1.bn.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.0.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.0.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.0.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.0.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.0.3.conv.weight - torch.Size([64, 64, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.0.3.conv.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.0.3.bn.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.0.3.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.1.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.1.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.1.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.1.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.1.3.conv.weight - torch.Size([64, 64, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.1.3.conv.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.1.3.bn.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.1.3.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.2.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.2.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.2.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.2.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.2.4.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.2.4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.3.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.3.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.tcn1.branches.3.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.tcn1.branches.3.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.residual.conv.weight - torch.Size([256, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.residual.conv.bias - torch.Size([256]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.7.residual.bn.weight - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.7.residual.bn.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.0.conv1.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.0.conv1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.0.conv2.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.0.conv2.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.0.conv3.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.0.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.0.conv4.weight - torch.Size([256, 32, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.0.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.1.conv1.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.1.conv1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.1.conv2.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.1.conv2.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.1.conv3.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.1.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.1.conv4.weight - torch.Size([256, 32, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.1.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.2.conv1.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.2.conv1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.2.conv2.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.2.conv2.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.2.conv3.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.2.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.convs.2.conv4.weight - torch.Size([256, 32, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.gcn1.convs.2.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.gcn1.bn.weight - torch.Size([256]): ConstantInit: val=1e-06, bias=0 backbone.net.8.gcn1.bn.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.0.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.0.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.0.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.0.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.0.3.conv.weight - torch.Size([64, 64, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.0.3.conv.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.0.3.bn.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.0.3.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.1.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.1.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.1.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.1.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.1.3.conv.weight - torch.Size([64, 64, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.1.3.conv.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.1.3.bn.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.1.3.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.2.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.2.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.2.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.2.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.2.4.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.2.4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.3.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.3.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.8.tcn1.branches.3.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.8.tcn1.branches.3.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.A - torch.Size([3, 25, 25]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.alpha - torch.Size([1]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.0.conv1.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.0.conv1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.0.conv2.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.0.conv2.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.0.conv3.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.0.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.0.conv4.weight - torch.Size([256, 32, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.0.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.1.conv1.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.1.conv1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.1.conv2.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.1.conv2.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.1.conv3.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.1.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.1.conv4.weight - torch.Size([256, 32, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.1.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.2.conv1.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.2.conv1.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.2.conv2.weight - torch.Size([32, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.2.conv2.bias - torch.Size([32]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.2.conv3.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.2.conv3.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.convs.2.conv4.weight - torch.Size([256, 32, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.gcn1.convs.2.conv4.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.gcn1.bn.weight - torch.Size([256]): ConstantInit: val=1e-06, bias=0 backbone.net.9.gcn1.bn.bias - torch.Size([256]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.0.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.0.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.0.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.0.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.0.3.conv.weight - torch.Size([64, 64, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.0.3.conv.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.0.3.bn.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.0.3.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.1.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.1.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.1.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.1.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.1.3.conv.weight - torch.Size([64, 64, 5, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.1.3.conv.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.1.3.bn.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.1.3.bn.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.2.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.2.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.2.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.2.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.2.4.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.2.4.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.3.0.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.3.0.bias - torch.Size([64]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0 backbone.net.9.tcn1.branches.3.1.weight - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN backbone.net.9.tcn1.branches.3.1.bias - torch.Size([64]): The value is the same before and after calling `init_weights` of RecognizerGCN cls_head.fc.weight - torch.Size([60, 256]): NormalInit: mean=0, std=0.01, bias=0 cls_head.fc.bias - torch.Size([60]): NormalInit: mean=0, std=0.01, bias=0 2023/03/06 23:23:17 - mmengine - INFO - Checkpoints will be saved to /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d. 2023/03/06 23:23:35 - mmengine - INFO - Epoch(train) [1][ 100/1567] lr: 9.9996e-02 eta: 1:14:22 time: 0.1323 data_time: 0.0070 memory: 3158 loss: 2.9710 top1_acc: 0.2500 top5_acc: 0.6250 loss_cls: 2.9710 2023/03/06 23:23:48 - mmengine - INFO - Epoch(train) [1][ 200/1567] lr: 9.9984e-02 eta: 1:03:11 time: 0.1252 data_time: 0.0064 memory: 3158 loss: 2.4077 top1_acc: 0.3750 top5_acc: 0.6875 loss_cls: 2.4077 2023/03/06 23:24:01 - mmengine - INFO - Epoch(train) [1][ 300/1567] lr: 9.9965e-02 eta: 0:59:32 time: 0.1291 data_time: 0.0069 memory: 3158 loss: 2.0504 top1_acc: 0.5000 top5_acc: 0.8750 loss_cls: 2.0504 2023/03/06 23:24:13 - mmengine - INFO - Epoch(train) [1][ 400/1567] lr: 9.9938e-02 eta: 0:57:34 time: 0.1264 data_time: 0.0066 memory: 3158 loss: 1.8165 top1_acc: 0.6875 top5_acc: 0.8750 loss_cls: 1.8165 2023/03/06 23:24:26 - mmengine - INFO - Epoch(train) [1][ 500/1567] lr: 9.9902e-02 eta: 0:56:26 time: 0.1285 data_time: 0.0064 memory: 3158 loss: 1.6022 top1_acc: 0.5000 top5_acc: 0.9375 loss_cls: 1.6022 2023/03/06 23:24:39 - mmengine - INFO - Epoch(train) [1][ 600/1567] lr: 9.9859e-02 eta: 0:55:28 time: 0.1289 data_time: 0.0080 memory: 3158 loss: 1.4758 top1_acc: 0.5000 top5_acc: 0.8750 loss_cls: 1.4758 2023/03/06 23:24:52 - mmengine - INFO - Epoch(train) [1][ 700/1567] lr: 9.9808e-02 eta: 0:54:53 time: 0.1317 data_time: 0.0066 memory: 3158 loss: 1.3330 top1_acc: 0.6250 top5_acc: 0.8750 loss_cls: 1.3330 2023/03/06 23:25:05 - mmengine - INFO - Epoch(train) [1][ 800/1567] lr: 9.9750e-02 eta: 0:54:22 time: 0.1283 data_time: 0.0065 memory: 3158 loss: 1.4036 top1_acc: 0.6875 top5_acc: 0.9375 loss_cls: 1.4036 2023/03/06 23:25:17 - mmengine - INFO - Epoch(train) [1][ 900/1567] lr: 9.9683e-02 eta: 0:53:47 time: 0.1254 data_time: 0.0064 memory: 3158 loss: 1.1917 top1_acc: 0.6250 top5_acc: 0.8125 loss_cls: 1.1917 2023/03/06 23:25:30 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:25:30 - mmengine - INFO - Epoch(train) [1][1000/1567] lr: 9.9609e-02 eta: 0:53:18 time: 0.1293 data_time: 0.0064 memory: 3158 loss: 0.9495 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.9495 2023/03/06 23:25:43 - mmengine - INFO - Epoch(train) [1][1100/1567] lr: 9.9527e-02 eta: 0:52:55 time: 0.1248 data_time: 0.0064 memory: 3158 loss: 1.0694 top1_acc: 0.6250 top5_acc: 0.8750 loss_cls: 1.0694 2023/03/06 23:25:56 - mmengine - INFO - Epoch(train) [1][1200/1567] lr: 9.9437e-02 eta: 0:52:36 time: 0.1336 data_time: 0.0071 memory: 3158 loss: 1.0414 top1_acc: 0.6875 top5_acc: 0.9375 loss_cls: 1.0414 2023/03/06 23:26:09 - mmengine - INFO - Epoch(train) [1][1300/1567] lr: 9.9339e-02 eta: 0:52:21 time: 0.1240 data_time: 0.0067 memory: 3158 loss: 1.0533 top1_acc: 0.6875 top5_acc: 0.8750 loss_cls: 1.0533 2023/03/06 23:26:22 - mmengine - INFO - Epoch(train) [1][1400/1567] lr: 9.9234e-02 eta: 0:52:10 time: 0.1357 data_time: 0.0065 memory: 3158 loss: 0.9680 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.9680 2023/03/06 23:26:36 - mmengine - INFO - Epoch(train) [1][1500/1567] lr: 9.9121e-02 eta: 0:51:59 time: 0.1305 data_time: 0.0066 memory: 3158 loss: 0.9409 top1_acc: 0.3750 top5_acc: 0.6875 loss_cls: 0.9409 2023/03/06 23:26:44 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:26:44 - mmengine - INFO - Epoch(train) [1][1567/1567] lr: 9.9040e-02 eta: 0:51:48 time: 0.1276 data_time: 0.0063 memory: 3158 loss: 1.0120 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 1.0120 2023/03/06 23:26:44 - mmengine - INFO - Saving checkpoint at 1 epochs 2023/03/06 23:26:49 - mmengine - INFO - Epoch(val) [1][100/129] eta: 0:00:01 time: 0.0377 data_time: 0.0078 memory: 223 2023/03/06 23:26:51 - mmengine - INFO - Epoch(val) [1][129/129] acc/top1: 0.3707 acc/top5: 0.7786 acc/mean1: 0.3707 2023/03/06 23:26:51 - mmengine - INFO - The best checkpoint with 0.3707 acc/top1 at 1 epoch is saved to best_acc/top1_epoch_1.pth. 2023/03/06 23:27:04 - mmengine - INFO - Epoch(train) [2][ 100/1567] lr: 9.8914e-02 eta: 0:51:35 time: 0.1317 data_time: 0.0067 memory: 3158 loss: 0.7672 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.7672 2023/03/06 23:27:17 - mmengine - INFO - Epoch(train) [2][ 200/1567] lr: 9.8781e-02 eta: 0:51:17 time: 0.1257 data_time: 0.0069 memory: 3158 loss: 0.8012 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.8012 2023/03/06 23:27:30 - mmengine - INFO - Epoch(train) [2][ 300/1567] lr: 9.8639e-02 eta: 0:51:01 time: 0.1278 data_time: 0.0068 memory: 3158 loss: 0.7625 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.7625 2023/03/06 23:27:43 - mmengine - INFO - Epoch(train) [2][ 400/1567] lr: 9.8491e-02 eta: 0:50:44 time: 0.1332 data_time: 0.0068 memory: 3158 loss: 0.6574 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.6574 2023/03/06 23:27:47 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:27:56 - mmengine - INFO - Epoch(train) [2][ 500/1567] lr: 9.8334e-02 eta: 0:50:25 time: 0.1249 data_time: 0.0069 memory: 3158 loss: 0.8187 top1_acc: 0.7500 top5_acc: 0.9375 loss_cls: 0.8187 2023/03/06 23:28:09 - mmengine - INFO - Epoch(train) [2][ 600/1567] lr: 9.8170e-02 eta: 0:50:11 time: 0.1307 data_time: 0.0070 memory: 3158 loss: 0.8061 top1_acc: 0.6875 top5_acc: 0.9375 loss_cls: 0.8061 2023/03/06 23:28:22 - mmengine - INFO - Epoch(train) [2][ 700/1567] lr: 9.7998e-02 eta: 0:49:57 time: 0.1309 data_time: 0.0067 memory: 3158 loss: 0.7958 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.7958 2023/03/06 23:28:35 - mmengine - INFO - Epoch(train) [2][ 800/1567] lr: 9.7819e-02 eta: 0:49:41 time: 0.1280 data_time: 0.0068 memory: 3158 loss: 0.7938 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.7938 2023/03/06 23:28:47 - mmengine - INFO - Epoch(train) [2][ 900/1567] lr: 9.7632e-02 eta: 0:49:24 time: 0.1286 data_time: 0.0069 memory: 3158 loss: 0.6454 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.6454 2023/03/06 23:29:01 - mmengine - INFO - Epoch(train) [2][1000/1567] lr: 9.7438e-02 eta: 0:49:13 time: 0.1326 data_time: 0.0072 memory: 3158 loss: 0.6147 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.6147 2023/03/06 23:29:13 - mmengine - INFO - Epoch(train) [2][1100/1567] lr: 9.7236e-02 eta: 0:48:55 time: 0.1247 data_time: 0.0068 memory: 3158 loss: 0.5740 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.5740 2023/03/06 23:29:26 - mmengine - INFO - Epoch(train) [2][1200/1567] lr: 9.7027e-02 eta: 0:48:40 time: 0.1269 data_time: 0.0069 memory: 3158 loss: 0.8133 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.8133 2023/03/06 23:29:39 - mmengine - INFO - Epoch(train) [2][1300/1567] lr: 9.6810e-02 eta: 0:48:24 time: 0.1258 data_time: 0.0070 memory: 3158 loss: 0.6209 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.6209 2023/03/06 23:29:52 - mmengine - INFO - Epoch(train) [2][1400/1567] lr: 9.6587e-02 eta: 0:48:12 time: 0.1360 data_time: 0.0068 memory: 3158 loss: 0.6251 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.6251 2023/03/06 23:29:56 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:30:05 - mmengine - INFO - Epoch(train) [2][1500/1567] lr: 9.6355e-02 eta: 0:47:58 time: 0.1280 data_time: 0.0071 memory: 3158 loss: 0.5752 top1_acc: 0.6875 top5_acc: 0.9375 loss_cls: 0.5752 2023/03/06 23:30:14 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:30:14 - mmengine - INFO - Epoch(train) [2][1567/1567] lr: 9.6196e-02 eta: 0:47:49 time: 0.1285 data_time: 0.0067 memory: 3158 loss: 0.9150 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.9150 2023/03/06 23:30:14 - mmengine - INFO - Saving checkpoint at 2 epochs 2023/03/06 23:30:18 - mmengine - INFO - Epoch(val) [2][100/129] eta: 0:00:01 time: 0.0361 data_time: 0.0074 memory: 223 2023/03/06 23:30:19 - mmengine - INFO - Epoch(val) [2][129/129] acc/top1: 0.6024 acc/top5: 0.9068 acc/mean1: 0.6023 2023/03/06 23:30:19 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_1.pth is removed 2023/03/06 23:30:20 - mmengine - INFO - The best checkpoint with 0.6024 acc/top1 at 2 epoch is saved to best_acc/top1_epoch_2.pth. 2023/03/06 23:30:32 - mmengine - INFO - Epoch(train) [3][ 100/1567] lr: 9.5953e-02 eta: 0:47:33 time: 0.1275 data_time: 0.0068 memory: 3158 loss: 0.6490 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.6490 2023/03/06 23:30:45 - mmengine - INFO - Epoch(train) [3][ 200/1567] lr: 9.5703e-02 eta: 0:47:19 time: 0.1286 data_time: 0.0068 memory: 3158 loss: 0.5952 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.5952 2023/03/06 23:30:58 - mmengine - INFO - Epoch(train) [3][ 300/1567] lr: 9.5445e-02 eta: 0:47:06 time: 0.1276 data_time: 0.0070 memory: 3158 loss: 0.5140 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.5140 2023/03/06 23:31:11 - mmengine - INFO - Epoch(train) [3][ 400/1567] lr: 9.5180e-02 eta: 0:46:51 time: 0.1257 data_time: 0.0067 memory: 3158 loss: 0.7440 top1_acc: 0.6875 top5_acc: 1.0000 loss_cls: 0.7440 2023/03/06 23:31:24 - mmengine - INFO - Epoch(train) [3][ 500/1567] lr: 9.4908e-02 eta: 0:46:35 time: 0.1252 data_time: 0.0067 memory: 3158 loss: 0.6553 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.6553 2023/03/06 23:31:36 - mmengine - INFO - Epoch(train) [3][ 600/1567] lr: 9.4629e-02 eta: 0:46:19 time: 0.1260 data_time: 0.0068 memory: 3158 loss: 0.5671 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.5671 2023/03/06 23:31:49 - mmengine - INFO - Epoch(train) [3][ 700/1567] lr: 9.4343e-02 eta: 0:46:05 time: 0.1246 data_time: 0.0067 memory: 3158 loss: 0.7042 top1_acc: 0.7500 top5_acc: 1.0000 loss_cls: 0.7042 2023/03/06 23:32:02 - mmengine - INFO - Epoch(train) [3][ 800/1567] lr: 9.4050e-02 eta: 0:45:50 time: 0.1270 data_time: 0.0070 memory: 3158 loss: 0.5807 top1_acc: 0.8125 top5_acc: 0.8750 loss_cls: 0.5807 2023/03/06 23:32:10 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:32:14 - mmengine - INFO - Epoch(train) [3][ 900/1567] lr: 9.3750e-02 eta: 0:45:35 time: 0.1244 data_time: 0.0067 memory: 3158 loss: 0.5888 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.5888 2023/03/06 23:32:27 - mmengine - INFO - Epoch(train) [3][1000/1567] lr: 9.3444e-02 eta: 0:45:22 time: 0.1319 data_time: 0.0069 memory: 3158 loss: 0.5718 top1_acc: 0.7500 top5_acc: 0.9375 loss_cls: 0.5718 2023/03/06 23:32:40 - mmengine - INFO - Epoch(train) [3][1100/1567] lr: 9.3130e-02 eta: 0:45:09 time: 0.1244 data_time: 0.0069 memory: 3158 loss: 0.4515 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.4515 2023/03/06 23:32:53 - mmengine - INFO - Epoch(train) [3][1200/1567] lr: 9.2810e-02 eta: 0:44:55 time: 0.1326 data_time: 0.0069 memory: 3158 loss: 0.6029 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.6029 2023/03/06 23:33:06 - mmengine - INFO - Epoch(train) [3][1300/1567] lr: 9.2483e-02 eta: 0:44:41 time: 0.1293 data_time: 0.0072 memory: 3158 loss: 0.5278 top1_acc: 0.7500 top5_acc: 0.9375 loss_cls: 0.5278 2023/03/06 23:33:19 - mmengine - INFO - Epoch(train) [3][1400/1567] lr: 9.2149e-02 eta: 0:44:27 time: 0.1320 data_time: 0.0070 memory: 3158 loss: 0.4881 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.4881 2023/03/06 23:33:32 - mmengine - INFO - Epoch(train) [3][1500/1567] lr: 9.1809e-02 eta: 0:44:14 time: 0.1384 data_time: 0.0070 memory: 3158 loss: 0.5170 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.5170 2023/03/06 23:33:41 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:33:41 - mmengine - INFO - Epoch(train) [3][1567/1567] lr: 9.1577e-02 eta: 0:44:07 time: 0.1286 data_time: 0.0068 memory: 3158 loss: 0.7181 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.7181 2023/03/06 23:33:41 - mmengine - INFO - Saving checkpoint at 3 epochs 2023/03/06 23:33:45 - mmengine - INFO - Epoch(val) [3][100/129] eta: 0:00:01 time: 0.0363 data_time: 0.0070 memory: 223 2023/03/06 23:33:46 - mmengine - INFO - Epoch(val) [3][129/129] acc/top1: 0.6522 acc/top5: 0.9224 acc/mean1: 0.6520 2023/03/06 23:33:46 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_2.pth is removed 2023/03/06 23:33:47 - mmengine - INFO - The best checkpoint with 0.6522 acc/top1 at 3 epoch is saved to best_acc/top1_epoch_3.pth. 2023/03/06 23:34:00 - mmengine - INFO - Epoch(train) [4][ 100/1567] lr: 9.1226e-02 eta: 0:43:54 time: 0.1318 data_time: 0.0069 memory: 3158 loss: 0.4382 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.4382 2023/03/06 23:34:13 - mmengine - INFO - Epoch(train) [4][ 200/1567] lr: 9.0868e-02 eta: 0:43:41 time: 0.1270 data_time: 0.0078 memory: 3158 loss: 0.4418 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.4418 2023/03/06 23:34:25 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:34:25 - mmengine - INFO - Epoch(train) [4][ 300/1567] lr: 9.0504e-02 eta: 0:43:27 time: 0.1278 data_time: 0.0069 memory: 3158 loss: 0.5110 top1_acc: 0.7500 top5_acc: 0.8750 loss_cls: 0.5110 2023/03/06 23:34:38 - mmengine - INFO - Epoch(train) [4][ 400/1567] lr: 9.0133e-02 eta: 0:43:13 time: 0.1281 data_time: 0.0069 memory: 3158 loss: 0.4928 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.4928 2023/03/06 23:34:52 - mmengine - INFO - Epoch(train) [4][ 500/1567] lr: 8.9756e-02 eta: 0:43:02 time: 0.1311 data_time: 0.0071 memory: 3158 loss: 0.5067 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.5067 2023/03/06 23:35:05 - mmengine - INFO - Epoch(train) [4][ 600/1567] lr: 8.9373e-02 eta: 0:42:49 time: 0.1317 data_time: 0.0069 memory: 3158 loss: 0.5311 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.5311 2023/03/06 23:35:18 - mmengine - INFO - Epoch(train) [4][ 700/1567] lr: 8.8984e-02 eta: 0:42:38 time: 0.1311 data_time: 0.0068 memory: 3158 loss: 0.5354 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.5354 2023/03/06 23:35:31 - mmengine - INFO - Epoch(train) [4][ 800/1567] lr: 8.8589e-02 eta: 0:42:24 time: 0.1272 data_time: 0.0068 memory: 3158 loss: 0.4916 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.4916 2023/03/06 23:35:44 - mmengine - INFO - Epoch(train) [4][ 900/1567] lr: 8.8187e-02 eta: 0:42:10 time: 0.1274 data_time: 0.0068 memory: 3158 loss: 0.5456 top1_acc: 0.7500 top5_acc: 1.0000 loss_cls: 0.5456 2023/03/06 23:35:56 - mmengine - INFO - Epoch(train) [4][1000/1567] lr: 8.7780e-02 eta: 0:41:56 time: 0.1239 data_time: 0.0068 memory: 3158 loss: 0.4726 top1_acc: 0.7500 top5_acc: 1.0000 loss_cls: 0.4726 2023/03/06 23:36:09 - mmengine - INFO - Epoch(train) [4][1100/1567] lr: 8.7367e-02 eta: 0:41:41 time: 0.1291 data_time: 0.0068 memory: 3158 loss: 0.4861 top1_acc: 0.7500 top5_acc: 1.0000 loss_cls: 0.4861 2023/03/06 23:36:22 - mmengine - INFO - Epoch(train) [4][1200/1567] lr: 8.6947e-02 eta: 0:41:29 time: 0.1345 data_time: 0.0069 memory: 3158 loss: 0.4354 top1_acc: 0.7500 top5_acc: 1.0000 loss_cls: 0.4354 2023/03/06 23:36:35 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:36:35 - mmengine - INFO - Epoch(train) [4][1300/1567] lr: 8.6522e-02 eta: 0:41:15 time: 0.1302 data_time: 0.0069 memory: 3158 loss: 0.4529 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.4529 2023/03/06 23:36:47 - mmengine - INFO - Epoch(train) [4][1400/1567] lr: 8.6092e-02 eta: 0:41:01 time: 0.1316 data_time: 0.0069 memory: 3158 loss: 0.5699 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.5699 2023/03/06 23:37:01 - mmengine - INFO - Epoch(train) [4][1500/1567] lr: 8.5655e-02 eta: 0:40:49 time: 0.1360 data_time: 0.0068 memory: 3158 loss: 0.4301 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.4301 2023/03/06 23:37:10 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:37:10 - mmengine - INFO - Epoch(train) [4][1567/1567] lr: 8.5360e-02 eta: 0:40:41 time: 0.1229 data_time: 0.0066 memory: 3158 loss: 0.6138 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.6138 2023/03/06 23:37:10 - mmengine - INFO - Saving checkpoint at 4 epochs 2023/03/06 23:37:14 - mmengine - INFO - Epoch(val) [4][100/129] eta: 0:00:01 time: 0.0362 data_time: 0.0077 memory: 223 2023/03/06 23:37:15 - mmengine - INFO - Epoch(val) [4][129/129] acc/top1: 0.7175 acc/top5: 0.9383 acc/mean1: 0.7172 2023/03/06 23:37:15 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_3.pth is removed 2023/03/06 23:37:16 - mmengine - INFO - The best checkpoint with 0.7175 acc/top1 at 4 epoch is saved to best_acc/top1_epoch_4.pth. 2023/03/06 23:37:28 - mmengine - INFO - Epoch(train) [5][ 100/1567] lr: 8.4914e-02 eta: 0:40:28 time: 0.1265 data_time: 0.0068 memory: 3158 loss: 0.4854 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.4854 2023/03/06 23:37:41 - mmengine - INFO - Epoch(train) [5][ 200/1567] lr: 8.4463e-02 eta: 0:40:14 time: 0.1255 data_time: 0.0069 memory: 3158 loss: 0.3300 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3300 2023/03/06 23:37:54 - mmengine - INFO - Epoch(train) [5][ 300/1567] lr: 8.4006e-02 eta: 0:40:00 time: 0.1319 data_time: 0.0073 memory: 3158 loss: 0.5270 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.5270 2023/03/06 23:38:07 - mmengine - INFO - Epoch(train) [5][ 400/1567] lr: 8.3544e-02 eta: 0:39:47 time: 0.1289 data_time: 0.0071 memory: 3158 loss: 0.4863 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.4863 2023/03/06 23:38:19 - mmengine - INFO - Epoch(train) [5][ 500/1567] lr: 8.3077e-02 eta: 0:39:33 time: 0.1282 data_time: 0.0069 memory: 3158 loss: 0.3407 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3407 2023/03/06 23:38:32 - mmengine - INFO - Epoch(train) [5][ 600/1567] lr: 8.2605e-02 eta: 0:39:20 time: 0.1281 data_time: 0.0069 memory: 3158 loss: 0.4589 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.4589 2023/03/06 23:38:45 - mmengine - INFO - Epoch(train) [5][ 700/1567] lr: 8.2127e-02 eta: 0:39:06 time: 0.1240 data_time: 0.0072 memory: 3158 loss: 0.4108 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.4108 2023/03/06 23:38:49 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:38:58 - mmengine - INFO - Epoch(train) [5][ 800/1567] lr: 8.1645e-02 eta: 0:38:53 time: 0.1304 data_time: 0.0067 memory: 3158 loss: 0.4803 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.4803 2023/03/06 23:39:11 - mmengine - INFO - Epoch(train) [5][ 900/1567] lr: 8.1157e-02 eta: 0:38:40 time: 0.1287 data_time: 0.0067 memory: 3158 loss: 0.4264 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.4264 2023/03/06 23:39:24 - mmengine - INFO - Epoch(train) [5][1000/1567] lr: 8.0665e-02 eta: 0:38:28 time: 0.1297 data_time: 0.0076 memory: 3158 loss: 0.3817 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3817 2023/03/06 23:39:37 - mmengine - INFO - Epoch(train) [5][1100/1567] lr: 8.0167e-02 eta: 0:38:15 time: 0.1284 data_time: 0.0067 memory: 3158 loss: 0.3295 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.3295 2023/03/06 23:39:50 - mmengine - INFO - Epoch(train) [5][1200/1567] lr: 7.9665e-02 eta: 0:38:03 time: 0.1305 data_time: 0.0067 memory: 3158 loss: 0.2924 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2924 2023/03/06 23:40:03 - mmengine - INFO - Epoch(train) [5][1300/1567] lr: 7.9159e-02 eta: 0:37:49 time: 0.1242 data_time: 0.0068 memory: 3158 loss: 0.3797 top1_acc: 0.7500 top5_acc: 1.0000 loss_cls: 0.3797 2023/03/06 23:40:16 - mmengine - INFO - Epoch(train) [5][1400/1567] lr: 7.8647e-02 eta: 0:37:36 time: 0.1262 data_time: 0.0075 memory: 3158 loss: 0.3645 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.3645 2023/03/06 23:40:29 - mmengine - INFO - Epoch(train) [5][1500/1567] lr: 7.8132e-02 eta: 0:37:23 time: 0.1281 data_time: 0.0069 memory: 3158 loss: 0.3461 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.3461 2023/03/06 23:40:37 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:40:37 - mmengine - INFO - Epoch(train) [5][1567/1567] lr: 7.7784e-02 eta: 0:37:14 time: 0.1244 data_time: 0.0066 memory: 3158 loss: 0.6115 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.6115 2023/03/06 23:40:37 - mmengine - INFO - Saving checkpoint at 5 epochs 2023/03/06 23:40:42 - mmengine - INFO - Epoch(val) [5][100/129] eta: 0:00:01 time: 0.0371 data_time: 0.0072 memory: 223 2023/03/06 23:40:43 - mmengine - INFO - Epoch(val) [5][129/129] acc/top1: 0.7359 acc/top5: 0.9467 acc/mean1: 0.7355 2023/03/06 23:40:43 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_4.pth is removed 2023/03/06 23:40:43 - mmengine - INFO - The best checkpoint with 0.7359 acc/top1 at 5 epoch is saved to best_acc/top1_epoch_5.pth. 2023/03/06 23:40:57 - mmengine - INFO - Epoch(train) [6][ 100/1567] lr: 7.7261e-02 eta: 0:37:02 time: 0.1330 data_time: 0.0071 memory: 3158 loss: 0.3690 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3690 2023/03/06 23:41:05 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:41:09 - mmengine - INFO - Epoch(train) [6][ 200/1567] lr: 7.6733e-02 eta: 0:36:48 time: 0.1265 data_time: 0.0069 memory: 3158 loss: 0.3864 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.3864 2023/03/06 23:41:22 - mmengine - INFO - Epoch(train) [6][ 300/1567] lr: 7.6202e-02 eta: 0:36:35 time: 0.1281 data_time: 0.0071 memory: 3158 loss: 0.3326 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.3326 2023/03/06 23:41:35 - mmengine - INFO - Epoch(train) [6][ 400/1567] lr: 7.5666e-02 eta: 0:36:21 time: 0.1263 data_time: 0.0069 memory: 3158 loss: 0.3549 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.3549 2023/03/06 23:41:48 - mmengine - INFO - Epoch(train) [6][ 500/1567] lr: 7.5126e-02 eta: 0:36:08 time: 0.1275 data_time: 0.0070 memory: 3158 loss: 0.4646 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.4646 2023/03/06 23:42:01 - mmengine - INFO - Epoch(train) [6][ 600/1567] lr: 7.4583e-02 eta: 0:35:56 time: 0.1298 data_time: 0.0069 memory: 3158 loss: 0.4916 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.4916 2023/03/06 23:42:14 - mmengine - INFO - Epoch(train) [6][ 700/1567] lr: 7.4035e-02 eta: 0:35:43 time: 0.1305 data_time: 0.0069 memory: 3158 loss: 0.3158 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.3158 2023/03/06 23:42:27 - mmengine - INFO - Epoch(train) [6][ 800/1567] lr: 7.3484e-02 eta: 0:35:31 time: 0.1325 data_time: 0.0067 memory: 3158 loss: 0.3918 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.3918 2023/03/06 23:42:41 - mmengine - INFO - Epoch(train) [6][ 900/1567] lr: 7.2929e-02 eta: 0:35:18 time: 0.1311 data_time: 0.0068 memory: 3158 loss: 0.4201 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.4201 2023/03/06 23:42:54 - mmengine - INFO - Epoch(train) [6][1000/1567] lr: 7.2371e-02 eta: 0:35:06 time: 0.1382 data_time: 0.0069 memory: 3158 loss: 0.3740 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.3740 2023/03/06 23:43:07 - mmengine - INFO - Epoch(train) [6][1100/1567] lr: 7.1809e-02 eta: 0:34:53 time: 0.1328 data_time: 0.0070 memory: 3158 loss: 0.3734 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.3734 2023/03/06 23:43:16 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:43:20 - mmengine - INFO - Epoch(train) [6][1200/1567] lr: 7.1243e-02 eta: 0:34:41 time: 0.1317 data_time: 0.0069 memory: 3158 loss: 0.3824 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.3824 2023/03/06 23:43:33 - mmengine - INFO - Epoch(train) [6][1300/1567] lr: 7.0674e-02 eta: 0:34:27 time: 0.1267 data_time: 0.0068 memory: 3158 loss: 0.3289 top1_acc: 0.6875 top5_acc: 1.0000 loss_cls: 0.3289 2023/03/06 23:43:45 - mmengine - INFO - Epoch(train) [6][1400/1567] lr: 7.0102e-02 eta: 0:34:13 time: 0.1237 data_time: 0.0068 memory: 3158 loss: 0.3218 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.3218 2023/03/06 23:43:58 - mmengine - INFO - Epoch(train) [6][1500/1567] lr: 6.9527e-02 eta: 0:34:00 time: 0.1307 data_time: 0.0068 memory: 3158 loss: 0.3541 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3541 2023/03/06 23:44:07 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:44:07 - mmengine - INFO - Epoch(train) [6][1567/1567] lr: 6.9140e-02 eta: 0:33:51 time: 0.1339 data_time: 0.0068 memory: 3158 loss: 0.5355 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.5355 2023/03/06 23:44:07 - mmengine - INFO - Saving checkpoint at 6 epochs 2023/03/06 23:44:11 - mmengine - INFO - Epoch(val) [6][100/129] eta: 0:00:01 time: 0.0364 data_time: 0.0077 memory: 223 2023/03/06 23:44:12 - mmengine - INFO - Epoch(val) [6][129/129] acc/top1: 0.8045 acc/top5: 0.9664 acc/mean1: 0.8044 2023/03/06 23:44:12 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_5.pth is removed 2023/03/06 23:44:13 - mmengine - INFO - The best checkpoint with 0.8045 acc/top1 at 6 epoch is saved to best_acc/top1_epoch_6.pth. 2023/03/06 23:44:25 - mmengine - INFO - Epoch(train) [7][ 100/1567] lr: 6.8560e-02 eta: 0:33:38 time: 0.1311 data_time: 0.0069 memory: 3158 loss: 0.3746 top1_acc: 0.7500 top5_acc: 0.9375 loss_cls: 0.3746 2023/03/06 23:44:38 - mmengine - INFO - Epoch(train) [7][ 200/1567] lr: 6.7976e-02 eta: 0:33:24 time: 0.1250 data_time: 0.0077 memory: 3158 loss: 0.3260 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.3260 2023/03/06 23:44:51 - mmengine - INFO - Epoch(train) [7][ 300/1567] lr: 6.7390e-02 eta: 0:33:11 time: 0.1311 data_time: 0.0070 memory: 3158 loss: 0.2793 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2793 2023/03/06 23:45:04 - mmengine - INFO - Epoch(train) [7][ 400/1567] lr: 6.6802e-02 eta: 0:32:59 time: 0.1295 data_time: 0.0069 memory: 3158 loss: 0.3003 top1_acc: 0.9375 top5_acc: 0.9375 loss_cls: 0.3003 2023/03/06 23:45:17 - mmengine - INFO - Epoch(train) [7][ 500/1567] lr: 6.6210e-02 eta: 0:32:46 time: 0.1277 data_time: 0.0074 memory: 3158 loss: 0.2744 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2744 2023/03/06 23:45:30 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:45:30 - mmengine - INFO - Epoch(train) [7][ 600/1567] lr: 6.5616e-02 eta: 0:32:33 time: 0.1245 data_time: 0.0070 memory: 3158 loss: 0.3731 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3731 2023/03/06 23:45:43 - mmengine - INFO - Epoch(train) [7][ 700/1567] lr: 6.5020e-02 eta: 0:32:19 time: 0.1245 data_time: 0.0069 memory: 3158 loss: 0.2531 top1_acc: 0.8125 top5_acc: 0.9375 loss_cls: 0.2531 2023/03/06 23:45:56 - mmengine - INFO - Epoch(train) [7][ 800/1567] lr: 6.4421e-02 eta: 0:32:06 time: 0.1315 data_time: 0.0068 memory: 3158 loss: 0.3962 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.3962 2023/03/06 23:46:09 - mmengine - INFO - Epoch(train) [7][ 900/1567] lr: 6.3820e-02 eta: 0:31:53 time: 0.1258 data_time: 0.0073 memory: 3158 loss: 0.3552 top1_acc: 0.6250 top5_acc: 0.9375 loss_cls: 0.3552 2023/03/06 23:46:21 - mmengine - INFO - Epoch(train) [7][1000/1567] lr: 6.3217e-02 eta: 0:31:40 time: 0.1341 data_time: 0.0070 memory: 3158 loss: 0.3027 top1_acc: 0.7500 top5_acc: 0.8750 loss_cls: 0.3027 2023/03/06 23:46:34 - mmengine - INFO - Epoch(train) [7][1100/1567] lr: 6.2612e-02 eta: 0:31:27 time: 0.1303 data_time: 0.0070 memory: 3158 loss: 0.2625 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2625 2023/03/06 23:46:47 - mmengine - INFO - Epoch(train) [7][1200/1567] lr: 6.2005e-02 eta: 0:31:14 time: 0.1275 data_time: 0.0069 memory: 3158 loss: 0.2828 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2828 2023/03/06 23:47:00 - mmengine - INFO - Epoch(train) [7][1300/1567] lr: 6.1396e-02 eta: 0:31:00 time: 0.1247 data_time: 0.0068 memory: 3158 loss: 0.2140 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2140 2023/03/06 23:47:12 - mmengine - INFO - Epoch(train) [7][1400/1567] lr: 6.0785e-02 eta: 0:30:47 time: 0.1240 data_time: 0.0069 memory: 3158 loss: 0.2461 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2461 2023/03/06 23:47:25 - mmengine - INFO - Epoch(train) [7][1500/1567] lr: 6.0172e-02 eta: 0:30:34 time: 0.1292 data_time: 0.0069 memory: 3158 loss: 0.3497 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.3497 2023/03/06 23:47:33 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:47:33 - mmengine - INFO - Epoch(train) [7][1567/1567] lr: 5.9761e-02 eta: 0:30:25 time: 0.1272 data_time: 0.0065 memory: 3158 loss: 0.5008 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.5008 2023/03/06 23:47:33 - mmengine - INFO - Saving checkpoint at 7 epochs 2023/03/06 23:47:38 - mmengine - INFO - Epoch(val) [7][100/129] eta: 0:00:01 time: 0.0360 data_time: 0.0075 memory: 223 2023/03/06 23:47:39 - mmengine - INFO - Epoch(val) [7][129/129] acc/top1: 0.7658 acc/top5: 0.9558 acc/mean1: 0.7657 2023/03/06 23:47:43 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:47:53 - mmengine - INFO - Epoch(train) [8][ 100/1567] lr: 5.9145e-02 eta: 0:30:13 time: 0.1276 data_time: 0.0067 memory: 3158 loss: 0.2666 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2666 2023/03/06 23:48:06 - mmengine - INFO - Epoch(train) [8][ 200/1567] lr: 5.8529e-02 eta: 0:30:00 time: 0.1361 data_time: 0.0068 memory: 3158 loss: 0.2592 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2592 2023/03/06 23:48:18 - mmengine - INFO - Epoch(train) [8][ 300/1567] lr: 5.7911e-02 eta: 0:29:46 time: 0.1240 data_time: 0.0067 memory: 3158 loss: 0.3407 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.3407 2023/03/06 23:48:31 - mmengine - INFO - Epoch(train) [8][ 400/1567] lr: 5.7292e-02 eta: 0:29:33 time: 0.1235 data_time: 0.0067 memory: 3158 loss: 0.2804 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2804 2023/03/06 23:48:44 - mmengine - INFO - Epoch(train) [8][ 500/1567] lr: 5.6671e-02 eta: 0:29:20 time: 0.1262 data_time: 0.0068 memory: 3158 loss: 0.2398 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2398 2023/03/06 23:48:56 - mmengine - INFO - Epoch(train) [8][ 600/1567] lr: 5.6050e-02 eta: 0:29:07 time: 0.1302 data_time: 0.0068 memory: 3158 loss: 0.2603 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2603 2023/03/06 23:49:10 - mmengine - INFO - Epoch(train) [8][ 700/1567] lr: 5.5427e-02 eta: 0:28:54 time: 0.1304 data_time: 0.0067 memory: 3158 loss: 0.4035 top1_acc: 0.8125 top5_acc: 1.0000 loss_cls: 0.4035 2023/03/06 23:49:22 - mmengine - INFO - Epoch(train) [8][ 800/1567] lr: 5.4804e-02 eta: 0:28:41 time: 0.1262 data_time: 0.0068 memory: 3158 loss: 0.3241 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.3241 2023/03/06 23:49:35 - mmengine - INFO - Epoch(train) [8][ 900/1567] lr: 5.4180e-02 eta: 0:28:28 time: 0.1302 data_time: 0.0069 memory: 3158 loss: 0.2795 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2795 2023/03/06 23:49:48 - mmengine - INFO - Epoch(train) [8][1000/1567] lr: 5.3556e-02 eta: 0:28:14 time: 0.1271 data_time: 0.0068 memory: 3158 loss: 0.2313 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2313 2023/03/06 23:49:52 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:50:00 - mmengine - INFO - Epoch(train) [8][1100/1567] lr: 5.2930e-02 eta: 0:28:01 time: 0.1257 data_time: 0.0067 memory: 3158 loss: 0.2683 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2683 2023/03/06 23:50:13 - mmengine - INFO - Epoch(train) [8][1200/1567] lr: 5.2305e-02 eta: 0:27:48 time: 0.1257 data_time: 0.0068 memory: 3158 loss: 0.2634 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2634 2023/03/06 23:50:26 - mmengine - INFO - Epoch(train) [8][1300/1567] lr: 5.1679e-02 eta: 0:27:35 time: 0.1322 data_time: 0.0068 memory: 3158 loss: 0.2701 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2701 2023/03/06 23:50:39 - mmengine - INFO - Epoch(train) [8][1400/1567] lr: 5.1052e-02 eta: 0:27:23 time: 0.1278 data_time: 0.0076 memory: 3158 loss: 0.2471 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2471 2023/03/06 23:50:52 - mmengine - INFO - Epoch(train) [8][1500/1567] lr: 5.0426e-02 eta: 0:27:09 time: 0.1280 data_time: 0.0073 memory: 3158 loss: 0.2213 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.2213 2023/03/06 23:51:00 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:51:00 - mmengine - INFO - Epoch(train) [8][1567/1567] lr: 5.0006e-02 eta: 0:27:00 time: 0.1225 data_time: 0.0066 memory: 3158 loss: 0.4490 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.4490 2023/03/06 23:51:00 - mmengine - INFO - Saving checkpoint at 8 epochs 2023/03/06 23:51:05 - mmengine - INFO - Epoch(val) [8][100/129] eta: 0:00:01 time: 0.0358 data_time: 0.0074 memory: 223 2023/03/06 23:51:06 - mmengine - INFO - Epoch(val) [8][129/129] acc/top1: 0.8086 acc/top5: 0.9688 acc/mean1: 0.8084 2023/03/06 23:51:06 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_6.pth is removed 2023/03/06 23:51:06 - mmengine - INFO - The best checkpoint with 0.8086 acc/top1 at 8 epoch is saved to best_acc/top1_epoch_8.pth. 2023/03/06 23:51:20 - mmengine - INFO - Epoch(train) [9][ 100/1567] lr: 4.9380e-02 eta: 0:26:48 time: 0.1300 data_time: 0.0069 memory: 3158 loss: 0.2307 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2307 2023/03/06 23:51:33 - mmengine - INFO - Epoch(train) [9][ 200/1567] lr: 4.8753e-02 eta: 0:26:35 time: 0.1328 data_time: 0.0068 memory: 3158 loss: 0.2137 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2137 2023/03/06 23:51:46 - mmengine - INFO - Epoch(train) [9][ 300/1567] lr: 4.8127e-02 eta: 0:26:23 time: 0.1313 data_time: 0.0069 memory: 3158 loss: 0.2195 top1_acc: 0.9375 top5_acc: 0.9375 loss_cls: 0.2195 2023/03/06 23:51:59 - mmengine - INFO - Epoch(train) [9][ 400/1567] lr: 4.7501e-02 eta: 0:26:10 time: 0.1274 data_time: 0.0072 memory: 3158 loss: 0.2935 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2935 2023/03/06 23:52:07 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:52:12 - mmengine - INFO - Epoch(train) [9][ 500/1567] lr: 4.6876e-02 eta: 0:25:57 time: 0.1362 data_time: 0.0076 memory: 3158 loss: 0.2596 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2596 2023/03/06 23:52:25 - mmengine - INFO - Epoch(train) [9][ 600/1567] lr: 4.6251e-02 eta: 0:25:44 time: 0.1289 data_time: 0.0069 memory: 3158 loss: 0.1973 top1_acc: 0.9375 top5_acc: 0.9375 loss_cls: 0.1973 2023/03/06 23:52:38 - mmengine - INFO - Epoch(train) [9][ 700/1567] lr: 4.5626e-02 eta: 0:25:31 time: 0.1271 data_time: 0.0068 memory: 3158 loss: 0.2267 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2267 2023/03/06 23:52:50 - mmengine - INFO - Epoch(train) [9][ 800/1567] lr: 4.5003e-02 eta: 0:25:17 time: 0.1310 data_time: 0.0067 memory: 3158 loss: 0.2209 top1_acc: 0.9375 top5_acc: 0.9375 loss_cls: 0.2209 2023/03/06 23:53:04 - mmengine - INFO - Epoch(train) [9][ 900/1567] lr: 4.4380e-02 eta: 0:25:05 time: 0.1323 data_time: 0.0068 memory: 3158 loss: 0.1607 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1607 2023/03/06 23:53:17 - mmengine - INFO - Epoch(train) [9][1000/1567] lr: 4.3757e-02 eta: 0:24:52 time: 0.1332 data_time: 0.0069 memory: 3158 loss: 0.2957 top1_acc: 0.8750 top5_acc: 0.9375 loss_cls: 0.2957 2023/03/06 23:53:30 - mmengine - INFO - Epoch(train) [9][1100/1567] lr: 4.3136e-02 eta: 0:24:39 time: 0.1240 data_time: 0.0069 memory: 3158 loss: 0.2182 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2182 2023/03/06 23:53:43 - mmengine - INFO - Epoch(train) [9][1200/1567] lr: 4.2516e-02 eta: 0:24:26 time: 0.1309 data_time: 0.0069 memory: 3158 loss: 0.2567 top1_acc: 0.9375 top5_acc: 0.9375 loss_cls: 0.2567 2023/03/06 23:53:55 - mmengine - INFO - Epoch(train) [9][1300/1567] lr: 4.1897e-02 eta: 0:24:13 time: 0.1267 data_time: 0.0068 memory: 3158 loss: 0.2398 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2398 2023/03/06 23:54:08 - mmengine - INFO - Epoch(train) [9][1400/1567] lr: 4.1280e-02 eta: 0:24:00 time: 0.1337 data_time: 0.0069 memory: 3158 loss: 0.2146 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2146 2023/03/06 23:54:17 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:54:22 - mmengine - INFO - Epoch(train) [9][1500/1567] lr: 4.0664e-02 eta: 0:23:48 time: 0.1319 data_time: 0.0068 memory: 3158 loss: 0.1920 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1920 2023/03/06 23:54:31 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:54:31 - mmengine - INFO - Epoch(train) [9][1567/1567] lr: 4.0252e-02 eta: 0:23:40 time: 0.1343 data_time: 0.0067 memory: 3158 loss: 0.3499 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.3499 2023/03/06 23:54:31 - mmengine - INFO - Saving checkpoint at 9 epochs 2023/03/06 23:54:35 - mmengine - INFO - Epoch(val) [9][100/129] eta: 0:00:01 time: 0.0360 data_time: 0.0071 memory: 223 2023/03/06 23:54:37 - mmengine - INFO - Epoch(val) [9][129/129] acc/top1: 0.8293 acc/top5: 0.9686 acc/mean1: 0.8293 2023/03/06 23:54:37 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_8.pth is removed 2023/03/06 23:54:37 - mmengine - INFO - The best checkpoint with 0.8293 acc/top1 at 9 epoch is saved to best_acc/top1_epoch_9.pth. 2023/03/06 23:54:50 - mmengine - INFO - Epoch(train) [10][ 100/1567] lr: 3.9638e-02 eta: 0:23:27 time: 0.1254 data_time: 0.0068 memory: 3158 loss: 0.2132 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.2132 2023/03/06 23:55:03 - mmengine - INFO - Epoch(train) [10][ 200/1567] lr: 3.9026e-02 eta: 0:23:14 time: 0.1316 data_time: 0.0069 memory: 3158 loss: 0.2334 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.2334 2023/03/06 23:55:16 - mmengine - INFO - Epoch(train) [10][ 300/1567] lr: 3.8415e-02 eta: 0:23:01 time: 0.1237 data_time: 0.0069 memory: 3158 loss: 0.1941 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1941 2023/03/06 23:55:29 - mmengine - INFO - Epoch(train) [10][ 400/1567] lr: 3.7807e-02 eta: 0:22:47 time: 0.1245 data_time: 0.0068 memory: 3158 loss: 0.1293 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1293 2023/03/06 23:55:41 - mmengine - INFO - Epoch(train) [10][ 500/1567] lr: 3.7200e-02 eta: 0:22:34 time: 0.1269 data_time: 0.0067 memory: 3158 loss: 0.1437 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1437 2023/03/06 23:55:54 - mmengine - INFO - Epoch(train) [10][ 600/1567] lr: 3.6596e-02 eta: 0:22:21 time: 0.1282 data_time: 0.0068 memory: 3158 loss: 0.1542 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1542 2023/03/06 23:56:07 - mmengine - INFO - Epoch(train) [10][ 700/1567] lr: 3.5993e-02 eta: 0:22:08 time: 0.1275 data_time: 0.0068 memory: 3158 loss: 0.2050 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2050 2023/03/06 23:56:20 - mmengine - INFO - Epoch(train) [10][ 800/1567] lr: 3.5393e-02 eta: 0:21:55 time: 0.1282 data_time: 0.0068 memory: 3158 loss: 0.2004 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.2004 2023/03/06 23:56:32 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:56:33 - mmengine - INFO - Epoch(train) [10][ 900/1567] lr: 3.4795e-02 eta: 0:21:42 time: 0.1397 data_time: 0.0084 memory: 3158 loss: 0.1742 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1742 2023/03/06 23:56:46 - mmengine - INFO - Epoch(train) [10][1000/1567] lr: 3.4199e-02 eta: 0:21:30 time: 0.1308 data_time: 0.0067 memory: 3158 loss: 0.1717 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1717 2023/03/06 23:56:58 - mmengine - INFO - Epoch(train) [10][1100/1567] lr: 3.3606e-02 eta: 0:21:16 time: 0.1239 data_time: 0.0068 memory: 3158 loss: 0.1919 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.1919 2023/03/06 23:57:11 - mmengine - INFO - Epoch(train) [10][1200/1567] lr: 3.3015e-02 eta: 0:21:03 time: 0.1236 data_time: 0.0066 memory: 3158 loss: 0.1315 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1315 2023/03/06 23:57:23 - mmengine - INFO - Epoch(train) [10][1300/1567] lr: 3.2428e-02 eta: 0:20:50 time: 0.1245 data_time: 0.0068 memory: 3158 loss: 0.1722 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1722 2023/03/06 23:57:35 - mmengine - INFO - Epoch(train) [10][1400/1567] lr: 3.1842e-02 eta: 0:20:36 time: 0.1241 data_time: 0.0068 memory: 3158 loss: 0.1425 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1425 2023/03/06 23:57:48 - mmengine - INFO - Epoch(train) [10][1500/1567] lr: 3.1260e-02 eta: 0:20:23 time: 0.1276 data_time: 0.0067 memory: 3158 loss: 0.1672 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.1672 2023/03/06 23:57:57 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:57:57 - mmengine - INFO - Epoch(train) [10][1567/1567] lr: 3.0872e-02 eta: 0:20:15 time: 0.1277 data_time: 0.0063 memory: 3158 loss: 0.2917 top1_acc: 0.0000 top5_acc: 1.0000 loss_cls: 0.2917 2023/03/06 23:57:57 - mmengine - INFO - Saving checkpoint at 10 epochs 2023/03/06 23:58:01 - mmengine - INFO - Epoch(val) [10][100/129] eta: 0:00:01 time: 0.0363 data_time: 0.0075 memory: 223 2023/03/06 23:58:02 - mmengine - INFO - Epoch(val) [10][129/129] acc/top1: 0.8374 acc/top5: 0.9702 acc/mean1: 0.8374 2023/03/06 23:58:02 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_9.pth is removed 2023/03/06 23:58:03 - mmengine - INFO - The best checkpoint with 0.8374 acc/top1 at 10 epoch is saved to best_acc/top1_epoch_10.pth. 2023/03/06 23:58:16 - mmengine - INFO - Epoch(train) [11][ 100/1567] lr: 3.0294e-02 eta: 0:20:02 time: 0.1308 data_time: 0.0068 memory: 3158 loss: 0.1720 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.1720 2023/03/06 23:58:28 - mmengine - INFO - Epoch(train) [11][ 200/1567] lr: 2.9720e-02 eta: 0:19:49 time: 0.1241 data_time: 0.0066 memory: 3158 loss: 0.1422 top1_acc: 0.9375 top5_acc: 0.9375 loss_cls: 0.1422 2023/03/06 23:58:41 - mmengine - INFO - Epoch(train) [11][ 300/1567] lr: 2.9149e-02 eta: 0:19:36 time: 0.1333 data_time: 0.0076 memory: 3158 loss: 0.1352 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1352 2023/03/06 23:58:45 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/06 23:58:55 - mmengine - INFO - Epoch(train) [11][ 400/1567] lr: 2.8581e-02 eta: 0:19:23 time: 0.1316 data_time: 0.0069 memory: 3158 loss: 0.1563 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.1563 2023/03/06 23:59:08 - mmengine - INFO - Epoch(train) [11][ 500/1567] lr: 2.8017e-02 eta: 0:19:10 time: 0.1306 data_time: 0.0071 memory: 3158 loss: 0.1416 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1416 2023/03/06 23:59:21 - mmengine - INFO - Epoch(train) [11][ 600/1567] lr: 2.7456e-02 eta: 0:18:58 time: 0.1355 data_time: 0.0069 memory: 3158 loss: 0.1286 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1286 2023/03/06 23:59:34 - mmengine - INFO - Epoch(train) [11][ 700/1567] lr: 2.6898e-02 eta: 0:18:45 time: 0.1324 data_time: 0.0067 memory: 3158 loss: 0.1144 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1144 2023/03/06 23:59:48 - mmengine - INFO - Epoch(train) [11][ 800/1567] lr: 2.6345e-02 eta: 0:18:32 time: 0.1379 data_time: 0.0069 memory: 3158 loss: 0.1556 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1556 2023/03/07 00:00:01 - mmengine - INFO - Epoch(train) [11][ 900/1567] lr: 2.5794e-02 eta: 0:18:19 time: 0.1304 data_time: 0.0073 memory: 3158 loss: 0.1223 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1223 2023/03/07 00:00:13 - mmengine - INFO - Epoch(train) [11][1000/1567] lr: 2.5248e-02 eta: 0:18:06 time: 0.1240 data_time: 0.0075 memory: 3158 loss: 0.1518 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1518 2023/03/07 00:00:26 - mmengine - INFO - Epoch(train) [11][1100/1567] lr: 2.4706e-02 eta: 0:17:53 time: 0.1327 data_time: 0.0069 memory: 3158 loss: 0.1349 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1349 2023/03/07 00:00:39 - mmengine - INFO - Epoch(train) [11][1200/1567] lr: 2.4167e-02 eta: 0:17:40 time: 0.1238 data_time: 0.0074 memory: 3158 loss: 0.0834 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0834 2023/03/07 00:00:52 - mmengine - INFO - Epoch(train) [11][1300/1567] lr: 2.3633e-02 eta: 0:17:27 time: 0.1279 data_time: 0.0074 memory: 3158 loss: 0.0902 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0902 2023/03/07 00:00:56 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:01:05 - mmengine - INFO - Epoch(train) [11][1400/1567] lr: 2.3103e-02 eta: 0:17:14 time: 0.1306 data_time: 0.0074 memory: 3158 loss: 0.1060 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1060 2023/03/07 00:01:18 - mmengine - INFO - Epoch(train) [11][1500/1567] lr: 2.2577e-02 eta: 0:17:01 time: 0.1265 data_time: 0.0070 memory: 3158 loss: 0.1157 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.1157 2023/03/07 00:01:26 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:01:26 - mmengine - INFO - Epoch(train) [11][1567/1567] lr: 2.2227e-02 eta: 0:16:53 time: 0.1250 data_time: 0.0067 memory: 3158 loss: 0.3740 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.3740 2023/03/07 00:01:26 - mmengine - INFO - Saving checkpoint at 11 epochs 2023/03/07 00:01:30 - mmengine - INFO - Epoch(val) [11][100/129] eta: 0:00:01 time: 0.0359 data_time: 0.0071 memory: 223 2023/03/07 00:01:32 - mmengine - INFO - Epoch(val) [11][129/129] acc/top1: 0.8409 acc/top5: 0.9682 acc/mean1: 0.8408 2023/03/07 00:01:32 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_10.pth is removed 2023/03/07 00:01:32 - mmengine - INFO - The best checkpoint with 0.8409 acc/top1 at 11 epoch is saved to best_acc/top1_epoch_11.pth. 2023/03/07 00:01:45 - mmengine - INFO - Epoch(train) [12][ 100/1567] lr: 2.1708e-02 eta: 0:16:40 time: 0.1265 data_time: 0.0079 memory: 3158 loss: 0.0948 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0948 2023/03/07 00:01:58 - mmengine - INFO - Epoch(train) [12][ 200/1567] lr: 2.1194e-02 eta: 0:16:27 time: 0.1300 data_time: 0.0068 memory: 3158 loss: 0.0802 top1_acc: 0.8750 top5_acc: 1.0000 loss_cls: 0.0802 2023/03/07 00:02:11 - mmengine - INFO - Epoch(train) [12][ 300/1567] lr: 2.0684e-02 eta: 0:16:14 time: 0.1344 data_time: 0.0068 memory: 3158 loss: 0.0926 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0926 2023/03/07 00:02:24 - mmengine - INFO - Epoch(train) [12][ 400/1567] lr: 2.0179e-02 eta: 0:16:01 time: 0.1302 data_time: 0.0075 memory: 3158 loss: 0.0980 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0980 2023/03/07 00:02:37 - mmengine - INFO - Epoch(train) [12][ 500/1567] lr: 1.9678e-02 eta: 0:15:48 time: 0.1316 data_time: 0.0069 memory: 3158 loss: 0.0778 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0778 2023/03/07 00:02:50 - mmengine - INFO - Epoch(train) [12][ 600/1567] lr: 1.9182e-02 eta: 0:15:35 time: 0.1337 data_time: 0.0068 memory: 3158 loss: 0.0636 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0636 2023/03/07 00:03:03 - mmengine - INFO - Epoch(train) [12][ 700/1567] lr: 1.8691e-02 eta: 0:15:22 time: 0.1305 data_time: 0.0067 memory: 3158 loss: 0.0536 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0536 2023/03/07 00:03:11 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:03:16 - mmengine - INFO - Epoch(train) [12][ 800/1567] lr: 1.8205e-02 eta: 0:15:09 time: 0.1242 data_time: 0.0068 memory: 3158 loss: 0.1244 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1244 2023/03/07 00:03:28 - mmengine - INFO - Epoch(train) [12][ 900/1567] lr: 1.7724e-02 eta: 0:14:56 time: 0.1246 data_time: 0.0067 memory: 3158 loss: 0.0671 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0671 2023/03/07 00:03:41 - mmengine - INFO - Epoch(train) [12][1000/1567] lr: 1.7248e-02 eta: 0:14:43 time: 0.1304 data_time: 0.0068 memory: 3158 loss: 0.0634 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0634 2023/03/07 00:03:55 - mmengine - INFO - Epoch(train) [12][1100/1567] lr: 1.6778e-02 eta: 0:14:30 time: 0.1246 data_time: 0.0073 memory: 3158 loss: 0.0654 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0654 2023/03/07 00:04:07 - mmengine - INFO - Epoch(train) [12][1200/1567] lr: 1.6312e-02 eta: 0:14:17 time: 0.1276 data_time: 0.0069 memory: 3158 loss: 0.0657 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0657 2023/03/07 00:04:20 - mmengine - INFO - Epoch(train) [12][1300/1567] lr: 1.5852e-02 eta: 0:14:04 time: 0.1278 data_time: 0.0068 memory: 3158 loss: 0.0608 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0608 2023/03/07 00:04:33 - mmengine - INFO - Epoch(train) [12][1400/1567] lr: 1.5397e-02 eta: 0:13:52 time: 0.1322 data_time: 0.0069 memory: 3158 loss: 0.0402 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0402 2023/03/07 00:04:46 - mmengine - INFO - Epoch(train) [12][1500/1567] lr: 1.4947e-02 eta: 0:13:39 time: 0.1242 data_time: 0.0070 memory: 3158 loss: 0.0576 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0576 2023/03/07 00:04:54 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:04:54 - mmengine - INFO - Epoch(train) [12][1567/1567] lr: 1.4649e-02 eta: 0:13:30 time: 0.1265 data_time: 0.0067 memory: 3158 loss: 0.2565 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.2565 2023/03/07 00:04:54 - mmengine - INFO - Saving checkpoint at 12 epochs 2023/03/07 00:04:59 - mmengine - INFO - Epoch(val) [12][100/129] eta: 0:00:01 time: 0.0358 data_time: 0.0073 memory: 223 2023/03/07 00:05:00 - mmengine - INFO - Epoch(val) [12][129/129] acc/top1: 0.8630 acc/top5: 0.9748 acc/mean1: 0.8630 2023/03/07 00:05:00 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_11.pth is removed 2023/03/07 00:05:00 - mmengine - INFO - The best checkpoint with 0.8630 acc/top1 at 12 epoch is saved to best_acc/top1_epoch_12.pth. 2023/03/07 00:05:14 - mmengine - INFO - Epoch(train) [13][ 100/1567] lr: 1.4209e-02 eta: 0:13:17 time: 0.1316 data_time: 0.0068 memory: 3158 loss: 0.0498 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0498 2023/03/07 00:05:26 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:05:27 - mmengine - INFO - Epoch(train) [13][ 200/1567] lr: 1.3774e-02 eta: 0:13:04 time: 0.1303 data_time: 0.0069 memory: 3158 loss: 0.0297 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0297 2023/03/07 00:05:40 - mmengine - INFO - Epoch(train) [13][ 300/1567] lr: 1.3345e-02 eta: 0:12:51 time: 0.1313 data_time: 0.0070 memory: 3158 loss: 0.0459 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0459 2023/03/07 00:05:53 - mmengine - INFO - Epoch(train) [13][ 400/1567] lr: 1.2922e-02 eta: 0:12:38 time: 0.1318 data_time: 0.0069 memory: 3158 loss: 0.0369 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0369 2023/03/07 00:06:06 - mmengine - INFO - Epoch(train) [13][ 500/1567] lr: 1.2505e-02 eta: 0:12:26 time: 0.1268 data_time: 0.0068 memory: 3158 loss: 0.0342 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0342 2023/03/07 00:06:19 - mmengine - INFO - Epoch(train) [13][ 600/1567] lr: 1.2093e-02 eta: 0:12:13 time: 0.1311 data_time: 0.0069 memory: 3158 loss: 0.0247 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0247 2023/03/07 00:06:32 - mmengine - INFO - Epoch(train) [13][ 700/1567] lr: 1.1687e-02 eta: 0:12:00 time: 0.1303 data_time: 0.0067 memory: 3158 loss: 0.0345 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0345 2023/03/07 00:06:45 - mmengine - INFO - Epoch(train) [13][ 800/1567] lr: 1.1288e-02 eta: 0:11:47 time: 0.1256 data_time: 0.0069 memory: 3158 loss: 0.0343 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0343 2023/03/07 00:06:58 - mmengine - INFO - Epoch(train) [13][ 900/1567] lr: 1.0894e-02 eta: 0:11:34 time: 0.1242 data_time: 0.0069 memory: 3158 loss: 0.0445 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0445 2023/03/07 00:07:11 - mmengine - INFO - Epoch(train) [13][1000/1567] lr: 1.0507e-02 eta: 0:11:21 time: 0.1309 data_time: 0.0069 memory: 3158 loss: 0.0371 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0371 2023/03/07 00:07:24 - mmengine - INFO - Epoch(train) [13][1100/1567] lr: 1.0126e-02 eta: 0:11:08 time: 0.1319 data_time: 0.0070 memory: 3158 loss: 0.0205 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0205 2023/03/07 00:07:37 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:07:37 - mmengine - INFO - Epoch(train) [13][1200/1567] lr: 9.7512e-03 eta: 0:10:55 time: 0.1282 data_time: 0.0070 memory: 3158 loss: 0.0258 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0258 2023/03/07 00:07:50 - mmengine - INFO - Epoch(train) [13][1300/1567] lr: 9.3826e-03 eta: 0:10:42 time: 0.1242 data_time: 0.0069 memory: 3158 loss: 0.0174 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0174 2023/03/07 00:08:02 - mmengine - INFO - Epoch(train) [13][1400/1567] lr: 9.0204e-03 eta: 0:10:29 time: 0.1245 data_time: 0.0078 memory: 3158 loss: 0.0500 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0500 2023/03/07 00:08:15 - mmengine - INFO - Epoch(train) [13][1500/1567] lr: 8.6647e-03 eta: 0:10:16 time: 0.1239 data_time: 0.0076 memory: 3158 loss: 0.0152 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0152 2023/03/07 00:08:23 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:08:23 - mmengine - INFO - Epoch(train) [13][1567/1567] lr: 8.4300e-03 eta: 0:10:07 time: 0.1251 data_time: 0.0067 memory: 3158 loss: 0.2042 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.2042 2023/03/07 00:08:23 - mmengine - INFO - Saving checkpoint at 13 epochs 2023/03/07 00:08:27 - mmengine - INFO - Epoch(val) [13][100/129] eta: 0:00:01 time: 0.0362 data_time: 0.0070 memory: 223 2023/03/07 00:08:28 - mmengine - INFO - Epoch(val) [13][129/129] acc/top1: 0.8794 acc/top5: 0.9788 acc/mean1: 0.8793 2023/03/07 00:08:28 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_12.pth is removed 2023/03/07 00:08:29 - mmengine - INFO - The best checkpoint with 0.8794 acc/top1 at 13 epoch is saved to best_acc/top1_epoch_13.pth. 2023/03/07 00:08:42 - mmengine - INFO - Epoch(train) [14][ 100/1567] lr: 8.0851e-03 eta: 0:09:54 time: 0.1245 data_time: 0.0068 memory: 3158 loss: 0.0194 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0194 2023/03/07 00:08:54 - mmengine - INFO - Epoch(train) [14][ 200/1567] lr: 7.7469e-03 eta: 0:09:41 time: 0.1244 data_time: 0.0070 memory: 3158 loss: 0.0274 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0274 2023/03/07 00:09:07 - mmengine - INFO - Epoch(train) [14][ 300/1567] lr: 7.4152e-03 eta: 0:09:28 time: 0.1292 data_time: 0.0069 memory: 3158 loss: 0.0175 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0175 2023/03/07 00:09:20 - mmengine - INFO - Epoch(train) [14][ 400/1567] lr: 7.0902e-03 eta: 0:09:15 time: 0.1366 data_time: 0.0068 memory: 3158 loss: 0.0235 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0235 2023/03/07 00:09:34 - mmengine - INFO - Epoch(train) [14][ 500/1567] lr: 6.7720e-03 eta: 0:09:03 time: 0.1358 data_time: 0.0068 memory: 3158 loss: 0.0151 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0151 2023/03/07 00:09:47 - mmengine - INFO - Epoch(train) [14][ 600/1567] lr: 6.4606e-03 eta: 0:08:50 time: 0.1306 data_time: 0.0069 memory: 3158 loss: 0.0182 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0182 2023/03/07 00:09:52 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:10:01 - mmengine - INFO - Epoch(train) [14][ 700/1567] lr: 6.1560e-03 eta: 0:08:37 time: 0.1310 data_time: 0.0067 memory: 3158 loss: 0.0190 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0190 2023/03/07 00:10:14 - mmengine - INFO - Epoch(train) [14][ 800/1567] lr: 5.8582e-03 eta: 0:08:24 time: 0.1313 data_time: 0.0069 memory: 3158 loss: 0.0251 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0251 2023/03/07 00:10:27 - mmengine - INFO - Epoch(train) [14][ 900/1567] lr: 5.5675e-03 eta: 0:08:11 time: 0.1254 data_time: 0.0068 memory: 3158 loss: 0.0178 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0178 2023/03/07 00:10:40 - mmengine - INFO - Epoch(train) [14][1000/1567] lr: 5.2836e-03 eta: 0:07:58 time: 0.1309 data_time: 0.0068 memory: 3158 loss: 0.0135 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0135 2023/03/07 00:10:53 - mmengine - INFO - Epoch(train) [14][1100/1567] lr: 5.0068e-03 eta: 0:07:45 time: 0.1310 data_time: 0.0069 memory: 3158 loss: 0.0164 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0164 2023/03/07 00:11:06 - mmengine - INFO - Epoch(train) [14][1200/1567] lr: 4.7371e-03 eta: 0:07:32 time: 0.1249 data_time: 0.0069 memory: 3158 loss: 0.0225 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0225 2023/03/07 00:11:18 - mmengine - INFO - Epoch(train) [14][1300/1567] lr: 4.4745e-03 eta: 0:07:19 time: 0.1238 data_time: 0.0068 memory: 3158 loss: 0.0078 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0078 2023/03/07 00:11:32 - mmengine - INFO - Epoch(train) [14][1400/1567] lr: 4.2190e-03 eta: 0:07:07 time: 0.1334 data_time: 0.0068 memory: 3158 loss: 0.0073 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0073 2023/03/07 00:11:45 - mmengine - INFO - Epoch(train) [14][1500/1567] lr: 3.9707e-03 eta: 0:06:54 time: 0.1319 data_time: 0.0069 memory: 3158 loss: 0.0119 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0119 2023/03/07 00:11:54 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:11:54 - mmengine - INFO - Epoch(train) [14][1567/1567] lr: 3.8084e-03 eta: 0:06:45 time: 0.1313 data_time: 0.0066 memory: 3158 loss: 0.1556 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.1556 2023/03/07 00:11:54 - mmengine - INFO - Saving checkpoint at 14 epochs 2023/03/07 00:11:58 - mmengine - INFO - Epoch(val) [14][100/129] eta: 0:00:01 time: 0.0359 data_time: 0.0074 memory: 223 2023/03/07 00:11:59 - mmengine - INFO - Epoch(val) [14][129/129] acc/top1: 0.8819 acc/top5: 0.9791 acc/mean1: 0.8818 2023/03/07 00:11:59 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_13.pth is removed 2023/03/07 00:12:00 - mmengine - INFO - The best checkpoint with 0.8819 acc/top1 at 14 epoch is saved to best_acc/top1_epoch_14.pth. 2023/03/07 00:12:08 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:12:13 - mmengine - INFO - Epoch(train) [15][ 100/1567] lr: 3.5722e-03 eta: 0:06:32 time: 0.1329 data_time: 0.0066 memory: 3158 loss: 0.0121 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0121 2023/03/07 00:12:26 - mmengine - INFO - Epoch(train) [15][ 200/1567] lr: 3.3433e-03 eta: 0:06:19 time: 0.1310 data_time: 0.0077 memory: 3158 loss: 0.0105 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0105 2023/03/07 00:12:39 - mmengine - INFO - Epoch(train) [15][ 300/1567] lr: 3.1217e-03 eta: 0:06:06 time: 0.1244 data_time: 0.0073 memory: 3158 loss: 0.0115 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0115 2023/03/07 00:12:52 - mmengine - INFO - Epoch(train) [15][ 400/1567] lr: 2.9075e-03 eta: 0:05:53 time: 0.1330 data_time: 0.0073 memory: 3158 loss: 0.0099 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0099 2023/03/07 00:13:05 - mmengine - INFO - Epoch(train) [15][ 500/1567] lr: 2.7007e-03 eta: 0:05:40 time: 0.1316 data_time: 0.0074 memory: 3158 loss: 0.0081 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0081 2023/03/07 00:13:18 - mmengine - INFO - Epoch(train) [15][ 600/1567] lr: 2.5013e-03 eta: 0:05:27 time: 0.1240 data_time: 0.0070 memory: 3158 loss: 0.0112 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0112 2023/03/07 00:13:31 - mmengine - INFO - Epoch(train) [15][ 700/1567] lr: 2.3093e-03 eta: 0:05:15 time: 0.1367 data_time: 0.0069 memory: 3158 loss: 0.0127 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0127 2023/03/07 00:13:44 - mmengine - INFO - Epoch(train) [15][ 800/1567] lr: 2.1249e-03 eta: 0:05:02 time: 0.1318 data_time: 0.0071 memory: 3158 loss: 0.0108 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0108 2023/03/07 00:13:57 - mmengine - INFO - Epoch(train) [15][ 900/1567] lr: 1.9479e-03 eta: 0:04:49 time: 0.1308 data_time: 0.0068 memory: 3158 loss: 0.0162 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0162 2023/03/07 00:14:10 - mmengine - INFO - Epoch(train) [15][1000/1567] lr: 1.7785e-03 eta: 0:04:36 time: 0.1285 data_time: 0.0069 memory: 3158 loss: 0.0099 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0099 2023/03/07 00:14:18 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:14:23 - mmengine - INFO - Epoch(train) [15][1100/1567] lr: 1.6167e-03 eta: 0:04:23 time: 0.1339 data_time: 0.0069 memory: 3158 loss: 0.0178 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0178 2023/03/07 00:14:36 - mmengine - INFO - Epoch(train) [15][1200/1567] lr: 1.4625e-03 eta: 0:04:10 time: 0.1270 data_time: 0.0070 memory: 3158 loss: 0.0089 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0089 2023/03/07 00:14:49 - mmengine - INFO - Epoch(train) [15][1300/1567] lr: 1.3159e-03 eta: 0:03:57 time: 0.1295 data_time: 0.0069 memory: 3158 loss: 0.0096 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0096 2023/03/07 00:15:02 - mmengine - INFO - Epoch(train) [15][1400/1567] lr: 1.1769e-03 eta: 0:03:44 time: 0.1265 data_time: 0.0070 memory: 3158 loss: 0.0136 top1_acc: 0.9375 top5_acc: 1.0000 loss_cls: 0.0136 2023/03/07 00:15:14 - mmengine - INFO - Epoch(train) [15][1500/1567] lr: 1.0456e-03 eta: 0:03:31 time: 0.1237 data_time: 0.0070 memory: 3158 loss: 0.0106 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0106 2023/03/07 00:15:23 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:15:23 - mmengine - INFO - Epoch(train) [15][1567/1567] lr: 9.6196e-04 eta: 0:03:22 time: 0.1333 data_time: 0.0067 memory: 3158 loss: 0.1580 top1_acc: 0.0000 top5_acc: 1.0000 loss_cls: 0.1580 2023/03/07 00:15:23 - mmengine - INFO - Saving checkpoint at 15 epochs 2023/03/07 00:15:27 - mmengine - INFO - Epoch(val) [15][100/129] eta: 0:00:01 time: 0.0370 data_time: 0.0086 memory: 223 2023/03/07 00:15:29 - mmengine - INFO - Epoch(val) [15][129/129] acc/top1: 0.8846 acc/top5: 0.9796 acc/mean1: 0.8845 2023/03/07 00:15:29 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_14.pth is removed 2023/03/07 00:15:29 - mmengine - INFO - The best checkpoint with 0.8846 acc/top1 at 15 epoch is saved to best_acc/top1_epoch_15.pth. 2023/03/07 00:15:42 - mmengine - INFO - Epoch(train) [16][ 100/1567] lr: 8.4351e-04 eta: 0:03:09 time: 0.1312 data_time: 0.0077 memory: 3158 loss: 0.0069 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0069 2023/03/07 00:15:56 - mmengine - INFO - Epoch(train) [16][ 200/1567] lr: 7.3277e-04 eta: 0:02:56 time: 0.1305 data_time: 0.0070 memory: 3158 loss: 0.0107 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0107 2023/03/07 00:16:09 - mmengine - INFO - Epoch(train) [16][ 300/1567] lr: 6.2978e-04 eta: 0:02:43 time: 0.1268 data_time: 0.0070 memory: 3158 loss: 0.0089 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0089 2023/03/07 00:16:21 - mmengine - INFO - Epoch(train) [16][ 400/1567] lr: 5.3453e-04 eta: 0:02:31 time: 0.1239 data_time: 0.0070 memory: 3158 loss: 0.0087 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0087 2023/03/07 00:16:33 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:16:34 - mmengine - INFO - Epoch(train) [16][ 500/1567] lr: 4.4705e-04 eta: 0:02:18 time: 0.1272 data_time: 0.0071 memory: 3158 loss: 0.0082 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0082 2023/03/07 00:16:47 - mmengine - INFO - Epoch(train) [16][ 600/1567] lr: 3.6735e-04 eta: 0:02:05 time: 0.1267 data_time: 0.0069 memory: 3158 loss: 0.0159 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0159 2023/03/07 00:16:59 - mmengine - INFO - Epoch(train) [16][ 700/1567] lr: 2.9544e-04 eta: 0:01:52 time: 0.1240 data_time: 0.0066 memory: 3158 loss: 0.0136 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0136 2023/03/07 00:17:12 - mmengine - INFO - Epoch(train) [16][ 800/1567] lr: 2.3134e-04 eta: 0:01:39 time: 0.1287 data_time: 0.0069 memory: 3158 loss: 0.0091 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0091 2023/03/07 00:17:25 - mmengine - INFO - Epoch(train) [16][ 900/1567] lr: 1.7505e-04 eta: 0:01:26 time: 0.1282 data_time: 0.0069 memory: 3158 loss: 0.0087 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0087 2023/03/07 00:17:38 - mmengine - INFO - Epoch(train) [16][1000/1567] lr: 1.2658e-04 eta: 0:01:13 time: 0.1305 data_time: 0.0069 memory: 3158 loss: 0.0069 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0069 2023/03/07 00:17:51 - mmengine - INFO - Epoch(train) [16][1100/1567] lr: 8.5947e-05 eta: 0:01:00 time: 0.1326 data_time: 0.0070 memory: 3158 loss: 0.0059 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0059 2023/03/07 00:18:04 - mmengine - INFO - Epoch(train) [16][1200/1567] lr: 5.3147e-05 eta: 0:00:47 time: 0.1255 data_time: 0.0068 memory: 3158 loss: 0.0075 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0075 2023/03/07 00:18:17 - mmengine - INFO - Epoch(train) [16][1300/1567] lr: 2.8190e-05 eta: 0:00:34 time: 0.1266 data_time: 0.0069 memory: 3158 loss: 0.0114 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0114 2023/03/07 00:18:30 - mmengine - INFO - Epoch(train) [16][1400/1567] lr: 1.1078e-05 eta: 0:00:21 time: 0.1259 data_time: 0.0069 memory: 3158 loss: 0.0069 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0069 2023/03/07 00:18:42 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:18:43 - mmengine - INFO - Epoch(train) [16][1500/1567] lr: 1.8150e-06 eta: 0:00:08 time: 0.1261 data_time: 0.0070 memory: 3158 loss: 0.0143 top1_acc: 1.0000 top5_acc: 1.0000 loss_cls: 0.0143 2023/03/07 00:18:51 - mmengine - INFO - Exp name: ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d_20230306_232252 2023/03/07 00:18:51 - mmengine - INFO - Epoch(train) [16][1567/1567] lr: 3.9252e-10 eta: 0:00:00 time: 0.1252 data_time: 0.0068 memory: 3158 loss: 0.2226 top1_acc: 0.0000 top5_acc: 0.0000 loss_cls: 0.2226 2023/03/07 00:18:51 - mmengine - INFO - Saving checkpoint at 16 epochs 2023/03/07 00:18:56 - mmengine - INFO - Epoch(val) [16][100/129] eta: 0:00:01 time: 0.0365 data_time: 0.0081 memory: 223 2023/03/07 00:18:57 - mmengine - INFO - Epoch(val) [16][129/129] acc/top1: 0.8855 acc/top5: 0.9797 acc/mean1: 0.8855 2023/03/07 00:18:57 - mmengine - INFO - The previous best checkpoint /mnt/petrelfs/daiwenxun/mmlab/mmaction2/projects/ctrgcn/work_dirs/ctrgcn_8xb16-joint-u100-80e_ntu60-xsub-keypoint-3d/best_acc/top1_epoch_15.pth is removed 2023/03/07 00:18:57 - mmengine - INFO - The best checkpoint with 0.8855 acc/top1 at 16 epoch is saved to best_acc/top1_epoch_16.pth.