2021-06-05 18:25:56,850 - mmaction - INFO - Environment info: ------------------------------------------------------------ sys.platform: linux Python: 3.7.6 (default, Jan 8 2020, 19:59:22) [GCC 7.3.0] CUDA available: True GPU 0,1,2,3,4,5,6,7: GeForce GTX 1080 Ti CUDA_HOME: /mnt/lustrenvme/share/polaris/dep/cuda-9.0-cudnn7.6.5 NVCC: Cuda compilation tools, release 9.0, V9.0.176 GCC: gcc (GCC) 5.4.0 PyTorch: 1.5.0 PyTorch compiling details: PyTorch built with: - GCC 5.4 - C++ Version: 201402 - Intel(R) Math Kernel Library Version 2020.0.1 Product Build 20200208 for Intel(R) 64 architecture applications - Intel(R) MKL-DNN v0.21.1 (Git Hash 912ce228837d1ce28e1a61806118835de03f5751) - OpenMP 201307 (a.k.a. OpenMP 4.0) - NNPACK is enabled - CPU capability usage: AVX2 - CUDA Runtime 9.0 - NVCC architecture flags: -gencode;arch=compute_52,code=sm_52;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_61,code=sm_61;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_70,code=compute_70 - CuDNN 7.6.5 - Magma 2.5.0 - Build settings: BLAS=MKL, BUILD_TYPE=Release, CXX_FLAGS= -Wno-deprecated -fvisibility-inlines-hidden -fopenmp -DNDEBUG -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DUSE_INTERNAL_THREADPOOL_IMPL -O2 -fPIC -Wno-narrowing -Wall -Wextra -Werror=return-type -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-sign-compare -Wno-unused-parameter -Wno-unused-variable -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-error=deprecated-declarations -Wno-error=pedantic -Wno-error=redundant-decls -Wno-error=old-style-cast -fdiagnostics-color=always -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, USE_CUDA=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=ON, USE_NCCL=ON, USE_NNPACK=ON, USE_OPENMP=ON, USE_STATIC_DISPATCH=OFF, TorchVision: 0.6.0 OpenCV: 4.2.0 MMCV: 1.3.1 MMCV Compiler: GCC 5.4 MMCV CUDA Compiler: 9.0 MMAction2: 0.15.0+985d284 ------------------------------------------------------------ 2021-06-05 18:25:56,852 - mmaction - INFO - Distributed training: True 2021-06-05 18:25:57,857 - mmaction - INFO - Config: model = dict( type='Recognizer2D', backbone=dict( type='ResNetTSM', pretrained='torchvision://resnet50', depth=50, norm_eval=False, shift_div=8, num_segments=8), cls_head=dict( type='TSMHead', num_classes=51, in_channels=2048, spatial_type='avg', consensus=dict(type='AvgConsensus', dim=1), dropout_ratio=0.5, init_std=0.001, is_shift=True, num_segments=8), train_cfg=None, test_cfg=dict(average_clips='prob')) optimizer = dict( type='SGD', constructor='TSMOptimizerConstructor', paramwise_cfg=dict(fc_lr5=True), lr=0.0015, momentum=0.9, weight_decay=0.0001) optimizer_config = dict(grad_clip=dict(max_norm=20, norm_type=2)) lr_config = dict(policy='step', step=[10, 20]) total_epochs = 25 checkpoint_config = dict(interval=1) log_config = dict(interval=20, hooks=[dict(type='TextLoggerHook')]) dist_params = dict(backend='nccl') log_level = 'INFO' load_from = 'https://download.openmmlab.com/mmaction/recognition/tsm/tsm_r50_256p_1x1x8_50e_kinetics400_rgb/tsm_r50_256p_1x1x8_50e_kinetics400_rgb_20200726-020785e2.pth' resume_from = None workflow = [('train', 1)] split = 1 dataset_type = 'RawframeDataset' data_root = 'data/hmdb51/rawframes' data_root_val = 'data/hmdb51/rawframes' ann_file_train = 'data/hmdb51/hmdb51_rgb_train_split_1.txt' ann_file_val = 'data/hmdb51/hmdb51_rgb_val_split_1.txt' ann_file_test = 'data/hmdb51/hmdb51_rgb_val_split_1.txt' img_norm_cfg = dict( mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False) mc_cfg = dict( server_list_cfg='/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3') train_pipeline = [ dict(type='SampleFrames', clip_len=1, frame_interval=1, num_clips=8), dict( type='RawFrameDecode', io_backend='memcached', server_list_cfg='/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3'), dict(type='Resize', scale=(-1, 256)), dict( type='MultiScaleCrop', input_size=224, scales=(1, 0.875, 0.75, 0.66), random_crop=False, max_wh_scale_gap=1, num_fixed_crops=13), dict(type='Resize', scale=(224, 224), keep_ratio=False), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False), dict(type='FormatShape', input_format='NCHW'), dict(type='Collect', keys=['imgs', 'label'], meta_keys=[]), dict(type='ToTensor', keys=['imgs', 'label']) ] val_pipeline = [ dict( type='SampleFrames', clip_len=1, frame_interval=1, num_clips=8, test_mode=True), dict( type='RawFrameDecode', io_backend='memcached', server_list_cfg='/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3'), dict(type='Resize', scale=(-1, 256)), dict(type='CenterCrop', crop_size=224), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False), dict(type='FormatShape', input_format='NCHW'), dict(type='Collect', keys=['imgs', 'label'], meta_keys=[]), dict(type='ToTensor', keys=['imgs']) ] test_pipeline = [ dict( type='SampleFrames', clip_len=1, frame_interval=1, num_clips=8, test_mode=True), dict( type='RawFrameDecode', io_backend='memcached', server_list_cfg='/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3'), dict(type='Resize', scale=(-1, 256)), dict(type='CenterCrop', crop_size=224), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False), dict(type='FormatShape', input_format='NCHW'), dict(type='Collect', keys=['imgs', 'label'], meta_keys=[]), dict(type='ToTensor', keys=['imgs']) ] data = dict( videos_per_gpu=12, workers_per_gpu=4, train=dict( type='RawframeDataset', ann_file='data/hmdb51/hmdb51_rgb_train_split_1.txt', data_prefix='data/hmdb51/rawframes', pipeline=[ dict( type='SampleFrames', clip_len=1, frame_interval=1, num_clips=8), dict( type='RawFrameDecode', io_backend='memcached', server_list_cfg= '/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3'), dict(type='Resize', scale=(-1, 256)), dict( type='MultiScaleCrop', input_size=224, scales=(1, 0.875, 0.75, 0.66), random_crop=False, max_wh_scale_gap=1, num_fixed_crops=13), dict(type='Resize', scale=(224, 224), keep_ratio=False), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False), dict(type='FormatShape', input_format='NCHW'), dict(type='Collect', keys=['imgs', 'label'], meta_keys=[]), dict(type='ToTensor', keys=['imgs', 'label']) ]), val=dict( type='RawframeDataset', ann_file='data/hmdb51/hmdb51_rgb_val_split_1.txt', data_prefix='data/hmdb51/rawframes', pipeline=[ dict( type='SampleFrames', clip_len=1, frame_interval=1, num_clips=8, test_mode=True), dict( type='RawFrameDecode', io_backend='memcached', server_list_cfg= '/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3'), dict(type='Resize', scale=(-1, 256)), dict(type='CenterCrop', crop_size=224), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False), dict(type='FormatShape', input_format='NCHW'), dict(type='Collect', keys=['imgs', 'label'], meta_keys=[]), dict(type='ToTensor', keys=['imgs']) ]), test=dict( type='RawframeDataset', ann_file='data/hmdb51/hmdb51_rgb_val_split_1.txt', data_prefix='data/hmdb51/rawframes', pipeline=[ dict( type='SampleFrames', clip_len=1, frame_interval=1, num_clips=8, test_mode=True), dict( type='RawFrameDecode', io_backend='memcached', server_list_cfg= '/mnt/lustre/share/memcached_client/server_list.conf', client_cfg='/mnt/lustre/share/memcached_client/client.conf', sys_path='/mnt/lustre/share/pymc/py3'), dict(type='Resize', scale=(-1, 256)), dict(type='CenterCrop', crop_size=224), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_bgr=False), dict(type='FormatShape', input_format='NCHW'), dict(type='Collect', keys=['imgs', 'label'], meta_keys=[]), dict(type='ToTensor', keys=['imgs']) ])) evaluation = dict( interval=1, metrics=['top_k_accuracy', 'mean_class_accuracy']) work_dir = './work_dirs/tsm_r50_1x1x8_25e_hmdb51_rgb/' gpu_ids = range(0, 8) omnisource = False module_hooks = [] 2021-06-05 18:26:15,727 - mmaction - INFO - These parameters in pretrained checkpoint are not loaded: {'fc.bias', 'fc.weight'} 2021-06-05 18:26:17,692 - mmaction - INFO - load checkpoint from https://download.openmmlab.com/mmaction/recognition/tsm/tsm_r50_256p_1x1x8_50e_kinetics400_rgb/tsm_r50_256p_1x1x8_50e_kinetics400_rgb_20200726-020785e2.pth 2021-06-05 18:26:17,692 - mmaction - INFO - Use load_from_http loader 2021-06-05 18:26:18,219 - mmaction - WARNING - The model and loaded state dict do not match exactly size mismatch for cls_head.fc_cls.weight: copying a param with shape torch.Size([400, 2048]) from checkpoint, the shape in current model is torch.Size([51, 2048]). size mismatch for cls_head.fc_cls.bias: copying a param with shape torch.Size([400]) from checkpoint, the shape in current model is torch.Size([51]). 2021-06-05 18:26:18,224 - mmaction - INFO - Start running, host: linjintao@SH-IDC1-10-5-30-99, work_dir: /mnt/lustre/linjintao/try/mmaction2_dev/work_dirs/tsm_r50_1x1x8_25e_hmdb51_rgb 2021-06-05 18:26:18,224 - mmaction - INFO - workflow: [('train', 1)], max: 25 epochs 2021-06-05 18:34:23,177 - mmaction - INFO - Epoch [1][20/38] lr: 1.500e-03, eta: 6:15:48, time: 24.245, data_time: 22.889, memory: 10388, top1_acc: 0.1365, top5_acc: 0.3417, loss_cls: 3.8020, loss: 3.8020, grad_norm: 1.5810 2021-06-05 18:35:00,702 - mmaction - INFO - Saving checkpoint at 1 epochs 2021-06-05 18:43:18,953 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 18:43:18,997 - mmaction - INFO - top1_acc 0.4484 top5_acc 0.7627 2021-06-05 18:43:18,997 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 18:43:19,058 - mmaction - INFO - mean_acc 0.4484 2021-06-05 18:43:19,915 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_1.pth. 2021-06-05 18:43:19,915 - mmaction - INFO - Best top1_acc is 0.4484 at 1 epoch. 2021-06-05 18:43:19,916 - mmaction - INFO - Epoch(val) [1][38] top1_acc: 0.4484, top5_acc: 0.7627, mean_class_accuracy: 0.4484 2021-06-05 18:51:10,064 - mmaction - INFO - Epoch [2][20/38] lr: 1.500e-03, eta: 4:04:47, time: 23.505, data_time: 22.419, memory: 10388, top1_acc: 0.4859, top5_acc: 0.7781, loss_cls: 2.7015, loss: 2.7015, grad_norm: 1.9811 2021-06-05 18:51:37,990 - mmaction - INFO - Saving checkpoint at 2 epochs 2021-06-05 18:59:03,221 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 18:59:03,265 - mmaction - INFO - top1_acc 0.6163 top5_acc 0.8693 2021-06-05 18:59:03,265 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 18:59:03,325 - mmaction - INFO - mean_acc 0.6163 2021-06-05 18:59:04,199 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_2.pth. 2021-06-05 18:59:04,199 - mmaction - INFO - Best top1_acc is 0.6163 at 2 epoch. 2021-06-05 18:59:04,200 - mmaction - INFO - Epoch(val) [2][38] top1_acc: 0.6163, top5_acc: 0.8693, mean_class_accuracy: 0.6163 2021-06-05 19:06:44,783 - mmaction - INFO - Epoch [3][20/38] lr: 1.500e-03, eta: 3:29:52, time: 23.026, data_time: 22.133, memory: 10388, top1_acc: 0.6115, top5_acc: 0.8682, loss_cls: 1.9186, loss: 1.9186, grad_norm: 2.0933 2021-06-05 19:07:05,351 - mmaction - INFO - Saving checkpoint at 3 epochs 2021-06-05 19:14:28,972 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 19:14:29,015 - mmaction - INFO - top1_acc 0.6458 top5_acc 0.9013 2021-06-05 19:14:29,016 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 19:14:29,076 - mmaction - INFO - mean_acc 0.6458 2021-06-05 19:14:29,939 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_3.pth. 2021-06-05 19:14:29,939 - mmaction - INFO - Best top1_acc is 0.6458 at 3 epoch. 2021-06-05 19:14:29,940 - mmaction - INFO - Epoch(val) [3][38] top1_acc: 0.6458, top5_acc: 0.9013, mean_class_accuracy: 0.6458 2021-06-05 19:22:05,800 - mmaction - INFO - Epoch [4][20/38] lr: 1.500e-03, eta: 3:09:55, time: 22.791, data_time: 21.888, memory: 10388, top1_acc: 0.6521, top5_acc: 0.8792, loss_cls: 1.5728, loss: 1.5728, grad_norm: 2.1967 2021-06-05 19:22:24,003 - mmaction - INFO - Saving checkpoint at 4 epochs 2021-06-05 19:30:03,232 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 19:30:03,279 - mmaction - INFO - top1_acc 0.6745 top5_acc 0.9078 2021-06-05 19:30:03,279 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 19:30:03,340 - mmaction - INFO - mean_acc 0.6745 2021-06-05 19:30:04,207 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_4.pth. 2021-06-05 19:30:04,207 - mmaction - INFO - Best top1_acc is 0.6745 at 4 epoch. 2021-06-05 19:30:04,208 - mmaction - INFO - Epoch(val) [4][38] top1_acc: 0.6745, top5_acc: 0.9078, mean_class_accuracy: 0.6745 2021-06-05 19:37:41,778 - mmaction - INFO - Epoch [5][20/38] lr: 1.500e-03, eta: 2:55:34, time: 22.877, data_time: 21.944, memory: 10388, top1_acc: 0.6682, top5_acc: 0.8990, loss_cls: 1.3969, loss: 1.3969, grad_norm: 2.3211 2021-06-05 19:38:00,039 - mmaction - INFO - Saving checkpoint at 5 epochs 2021-06-05 19:45:12,525 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 19:45:12,570 - mmaction - INFO - top1_acc 0.6869 top5_acc 0.9150 2021-06-05 19:45:12,570 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 19:45:12,630 - mmaction - INFO - mean_acc 0.6869 2021-06-05 19:45:13,493 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_5.pth. 2021-06-05 19:45:13,493 - mmaction - INFO - Best top1_acc is 0.6869 at 5 epoch. 2021-06-05 19:45:13,494 - mmaction - INFO - Epoch(val) [5][38] top1_acc: 0.6869, top5_acc: 0.9150, mean_class_accuracy: 0.6869 2021-06-05 19:52:39,004 - mmaction - INFO - Epoch [6][20/38] lr: 1.500e-03, eta: 2:42:56, time: 22.274, data_time: 21.340, memory: 10388, top1_acc: 0.6927, top5_acc: 0.9078, loss_cls: 1.2595, loss: 1.2595, grad_norm: 2.4563 2021-06-05 19:52:57,718 - mmaction - INFO - Saving checkpoint at 6 epochs 2021-06-05 20:00:12,908 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 20:00:12,956 - mmaction - INFO - top1_acc 0.6915 top5_acc 0.9144 2021-06-05 20:00:12,956 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 20:00:13,018 - mmaction - INFO - mean_acc 0.6915 2021-06-05 20:00:13,877 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_6.pth. 2021-06-05 20:00:13,877 - mmaction - INFO - Best top1_acc is 0.6915 at 6 epoch. 2021-06-05 20:00:13,878 - mmaction - INFO - Epoch(val) [6][38] top1_acc: 0.6915, top5_acc: 0.9144, mean_class_accuracy: 0.6915 2021-06-05 20:07:25,360 - mmaction - INFO - Epoch [7][20/38] lr: 1.500e-03, eta: 2:31:14, time: 21.572, data_time: 20.637, memory: 10388, top1_acc: 0.6969, top5_acc: 0.9089, loss_cls: 1.1855, loss: 1.1855, grad_norm: 2.4911 2021-06-05 20:07:43,856 - mmaction - INFO - Saving checkpoint at 7 epochs 2021-06-05 20:15:13,821 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 20:15:13,870 - mmaction - INFO - top1_acc 0.7052 top5_acc 0.9137 2021-06-05 20:15:13,870 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 20:15:13,936 - mmaction - INFO - mean_acc 0.7052 2021-06-05 20:15:14,796 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_7.pth. 2021-06-05 20:15:14,796 - mmaction - INFO - Best top1_acc is 0.7052 at 7 epoch. 2021-06-05 20:15:14,797 - mmaction - INFO - Epoch(val) [7][38] top1_acc: 0.7052, top5_acc: 0.9137, mean_class_accuracy: 0.7052 2021-06-05 20:22:30,573 - mmaction - INFO - Epoch [8][20/38] lr: 1.500e-03, eta: 2:20:54, time: 21.787, data_time: 20.843, memory: 10388, top1_acc: 0.7099, top5_acc: 0.9130, loss_cls: 1.1068, loss: 1.1068, grad_norm: 2.6262 2021-06-05 20:22:47,564 - mmaction - INFO - Saving checkpoint at 8 epochs 2021-06-05 20:30:23,955 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 20:30:23,998 - mmaction - INFO - top1_acc 0.7085 top5_acc 0.9261 2021-06-05 20:30:23,998 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 20:30:24,059 - mmaction - INFO - mean_acc 0.7085 2021-06-05 20:30:24,913 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_8.pth. 2021-06-05 20:30:24,914 - mmaction - INFO - Best top1_acc is 0.7085 at 8 epoch. 2021-06-05 20:30:24,915 - mmaction - INFO - Epoch(val) [8][38] top1_acc: 0.7085, top5_acc: 0.9261, mean_class_accuracy: 0.7085 2021-06-05 20:38:05,197 - mmaction - INFO - Epoch [9][20/38] lr: 1.500e-03, eta: 2:12:05, time: 23.012, data_time: 22.119, memory: 10388, top1_acc: 0.7281, top5_acc: 0.9214, loss_cls: 1.0497, loss: 1.0497, grad_norm: 2.6326 2021-06-05 20:38:23,892 - mmaction - INFO - Saving checkpoint at 9 epochs 2021-06-05 20:45:58,208 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 20:45:58,252 - mmaction - INFO - top1_acc 0.7176 top5_acc 0.9157 2021-06-05 20:45:58,252 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 20:45:58,312 - mmaction - INFO - mean_acc 0.7176 2021-06-05 20:45:59,171 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_9.pth. 2021-06-05 20:45:59,172 - mmaction - INFO - Best top1_acc is 0.7176 at 9 epoch. 2021-06-05 20:45:59,173 - mmaction - INFO - Epoch(val) [9][38] top1_acc: 0.7176, top5_acc: 0.9157, mean_class_accuracy: 0.7176 2021-06-05 20:53:44,103 - mmaction - INFO - Epoch [10][20/38] lr: 1.500e-03, eta: 2:03:37, time: 23.245, data_time: 22.307, memory: 10388, top1_acc: 0.7401, top5_acc: 0.9266, loss_cls: 0.9939, loss: 0.9939, grad_norm: 2.6957 2021-06-05 20:54:02,780 - mmaction - INFO - Saving checkpoint at 10 epochs 2021-06-05 21:01:28,096 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 21:01:28,140 - mmaction - INFO - top1_acc 0.7268 top5_acc 0.9203 2021-06-05 21:01:28,140 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 21:01:28,200 - mmaction - INFO - mean_acc 0.7268 2021-06-05 21:01:29,065 - mmaction - INFO - Now best checkpoint is saved as best_top1_acc_epoch_10.pth. 2021-06-05 21:01:29,065 - mmaction - INFO - Best top1_acc is 0.7268 at 10 epoch. 2021-06-05 21:01:29,066 - mmaction - INFO - Epoch(val) [10][38] top1_acc: 0.7268, top5_acc: 0.9203, mean_class_accuracy: 0.7268 2021-06-05 21:08:51,967 - mmaction - INFO - Epoch [11][20/38] lr: 1.500e-04, eta: 1:54:48, time: 22.143, data_time: 21.221, memory: 10388, top1_acc: 0.7719, top5_acc: 0.9302, loss_cls: 0.9231, loss: 0.9231, grad_norm: 2.6058 2021-06-05 21:09:10,751 - mmaction - INFO - Saving checkpoint at 11 epochs 2021-06-05 21:16:47,605 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 21:16:47,649 - mmaction - INFO - top1_acc 0.7196 top5_acc 0.9229 2021-06-05 21:16:47,649 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 21:16:47,709 - mmaction - INFO - mean_acc 0.7196 2021-06-05 21:16:47,710 - mmaction - INFO - Epoch(val) [11][38] top1_acc: 0.7196, top5_acc: 0.9229, mean_class_accuracy: 0.7196 2021-06-05 21:24:01,787 - mmaction - INFO - Epoch [12][20/38] lr: 1.500e-04, eta: 1:46:03, time: 21.700, data_time: 20.792, memory: 10388, top1_acc: 0.7698, top5_acc: 0.9354, loss_cls: 0.9210, loss: 0.9210, grad_norm: 2.6200 2021-06-05 21:24:18,798 - mmaction - INFO - Saving checkpoint at 12 epochs 2021-06-05 21:31:38,856 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 21:31:38,900 - mmaction - INFO - top1_acc 0.7196 top5_acc 0.9242 2021-06-05 21:31:38,900 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 21:31:38,960 - mmaction - INFO - mean_acc 0.7196 2021-06-05 21:31:38,961 - mmaction - INFO - Epoch(val) [12][38] top1_acc: 0.7196, top5_acc: 0.9242, mean_class_accuracy: 0.7196 2021-06-05 21:39:04,474 - mmaction - INFO - Epoch [13][20/38] lr: 1.500e-04, eta: 1:37:44, time: 22.275, data_time: 21.354, memory: 10388, top1_acc: 0.7521, top5_acc: 0.9307, loss_cls: 0.9168, loss: 0.9168, grad_norm: 2.6558 2021-06-05 21:39:21,411 - mmaction - INFO - Saving checkpoint at 13 epochs 2021-06-05 21:46:25,961 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 21:46:26,010 - mmaction - INFO - top1_acc 0.7216 top5_acc 0.9235 2021-06-05 21:46:26,010 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 21:46:26,075 - mmaction - INFO - mean_acc 0.7216 2021-06-05 21:46:26,076 - mmaction - INFO - Epoch(val) [13][38] top1_acc: 0.7216, top5_acc: 0.9235, mean_class_accuracy: 0.7216 2021-06-05 21:53:38,622 - mmaction - INFO - Epoch [14][20/38] lr: 1.500e-04, eta: 1:29:22, time: 21.625, data_time: 20.717, memory: 10388, top1_acc: 0.7484, top5_acc: 0.9292, loss_cls: 0.9165, loss: 0.9165, grad_norm: 2.6349 2021-06-05 21:53:55,397 - mmaction - INFO - Saving checkpoint at 14 epochs 2021-06-05 22:01:17,078 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 22:01:17,127 - mmaction - INFO - top1_acc 0.7209 top5_acc 0.9242 2021-06-05 22:01:17,127 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 22:01:17,192 - mmaction - INFO - mean_acc 0.7209 2021-06-05 22:01:17,193 - mmaction - INFO - Epoch(val) [14][38] top1_acc: 0.7209, top5_acc: 0.9242, mean_class_accuracy: 0.7209 2021-06-05 22:08:36,310 - mmaction - INFO - Epoch [15][20/38] lr: 1.500e-04, eta: 1:21:14, time: 21.953, data_time: 21.027, memory: 10388, top1_acc: 0.7698, top5_acc: 0.9401, loss_cls: 0.9035, loss: 0.9035, grad_norm: 2.6534 2021-06-05 22:08:55,082 - mmaction - INFO - Saving checkpoint at 15 epochs 2021-06-05 22:16:16,044 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 22:16:16,093 - mmaction - INFO - top1_acc 0.7216 top5_acc 0.9190 2021-06-05 22:16:16,093 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 22:16:16,158 - mmaction - INFO - mean_acc 0.7216 2021-06-05 22:16:16,159 - mmaction - INFO - Epoch(val) [15][38] top1_acc: 0.7216, top5_acc: 0.9190, mean_class_accuracy: 0.7216 2021-06-05 22:23:52,168 - mmaction - INFO - Epoch [16][20/38] lr: 1.500e-04, eta: 1:13:23, time: 22.798, data_time: 21.870, memory: 10388, top1_acc: 0.7604, top5_acc: 0.9370, loss_cls: 0.9063, loss: 0.9063, grad_norm: 2.7090 2021-06-05 22:24:10,814 - mmaction - INFO - Saving checkpoint at 16 epochs 2021-06-05 22:31:47,387 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 22:31:47,436 - mmaction - INFO - top1_acc 0.7183 top5_acc 0.9235 2021-06-05 22:31:47,436 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 22:31:47,501 - mmaction - INFO - mean_acc 0.7183 2021-06-05 22:31:47,502 - mmaction - INFO - Epoch(val) [16][38] top1_acc: 0.7183, top5_acc: 0.9235, mean_class_accuracy: 0.7183 2021-06-05 22:39:23,231 - mmaction - INFO - Epoch [17][20/38] lr: 1.500e-04, eta: 1:05:33, time: 22.784, data_time: 21.832, memory: 10388, top1_acc: 0.7464, top5_acc: 0.9318, loss_cls: 0.9597, loss: 0.9597, grad_norm: 2.7558 2021-06-05 22:39:41,435 - mmaction - INFO - Saving checkpoint at 17 epochs 2021-06-05 22:47:15,237 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 22:47:15,288 - mmaction - INFO - top1_acc 0.7268 top5_acc 0.9261 2021-06-05 22:47:15,288 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 22:47:15,354 - mmaction - INFO - mean_acc 0.7268 2021-06-05 22:47:15,354 - mmaction - INFO - Epoch(val) [17][38] top1_acc: 0.7268, top5_acc: 0.9261, mean_class_accuracy: 0.7268 2021-06-05 22:54:52,178 - mmaction - INFO - Epoch [18][20/38] lr: 1.500e-04, eta: 0:57:46, time: 22.839, data_time: 21.951, memory: 10388, top1_acc: 0.7734, top5_acc: 0.9297, loss_cls: 0.8987, loss: 0.8987, grad_norm: 2.7127 2021-06-05 22:55:10,845 - mmaction - INFO - Saving checkpoint at 18 epochs 2021-06-05 23:02:40,669 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 23:02:40,712 - mmaction - INFO - top1_acc 0.7248 top5_acc 0.9281 2021-06-05 23:02:40,712 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 23:02:40,768 - mmaction - INFO - mean_acc 0.7248 2021-06-05 23:02:40,768 - mmaction - INFO - Epoch(val) [18][38] top1_acc: 0.7248, top5_acc: 0.9281, mean_class_accuracy: 0.7248 2021-06-05 23:10:08,258 - mmaction - INFO - Epoch [19][20/38] lr: 1.500e-04, eta: 0:49:56, time: 22.372, data_time: 21.447, memory: 10388, top1_acc: 0.7521, top5_acc: 0.9344, loss_cls: 0.8969, loss: 0.8969, grad_norm: 2.7044 2021-06-05 23:10:26,980 - mmaction - INFO - Saving checkpoint at 19 epochs 2021-06-05 23:17:39,506 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 23:17:39,559 - mmaction - INFO - top1_acc 0.7248 top5_acc 0.9268 2021-06-05 23:17:39,559 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 23:17:39,624 - mmaction - INFO - mean_acc 0.7248 2021-06-05 23:17:39,625 - mmaction - INFO - Epoch(val) [19][38] top1_acc: 0.7248, top5_acc: 0.9268, mean_class_accuracy: 0.7248 2021-06-05 23:25:10,971 - mmaction - INFO - Epoch [20][20/38] lr: 1.500e-04, eta: 0:42:10, time: 22.565, data_time: 21.644, memory: 10388, top1_acc: 0.7615, top5_acc: 0.9406, loss_cls: 0.8863, loss: 0.8863, grad_norm: 2.6270 2021-06-05 23:25:29,761 - mmaction - INFO - Saving checkpoint at 20 epochs 2021-06-05 23:33:01,631 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 23:33:01,680 - mmaction - INFO - top1_acc 0.7203 top5_acc 0.9307 2021-06-05 23:33:01,680 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 23:33:01,745 - mmaction - INFO - mean_acc 0.7203 2021-06-05 23:33:01,746 - mmaction - INFO - Epoch(val) [20][38] top1_acc: 0.7203, top5_acc: 0.9307, mean_class_accuracy: 0.7203 2021-06-05 23:40:51,341 - mmaction - INFO - Epoch [21][20/38] lr: 1.500e-05, eta: 0:34:29, time: 23.476, data_time: 22.588, memory: 10388, top1_acc: 0.7714, top5_acc: 0.9359, loss_cls: 0.8918, loss: 0.8918, grad_norm: 2.6933 2021-06-05 23:41:09,856 - mmaction - INFO - Saving checkpoint at 21 epochs 2021-06-05 23:48:34,668 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-05 23:48:34,722 - mmaction - INFO - top1_acc 0.7248 top5_acc 0.9268 2021-06-05 23:48:34,722 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-05 23:48:34,787 - mmaction - INFO - mean_acc 0.7248 2021-06-05 23:48:34,788 - mmaction - INFO - Epoch(val) [21][38] top1_acc: 0.7248, top5_acc: 0.9268, mean_class_accuracy: 0.7248 2021-06-05 23:56:14,999 - mmaction - INFO - Epoch [22][20/38] lr: 1.500e-05, eta: 0:26:46, time: 23.008, data_time: 22.064, memory: 10388, top1_acc: 0.7734, top5_acc: 0.9406, loss_cls: 0.8825, loss: 0.8825, grad_norm: 2.6786 2021-06-05 23:56:33,155 - mmaction - INFO - Saving checkpoint at 22 epochs 2021-06-06 00:04:13,401 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-06 00:04:13,447 - mmaction - INFO - top1_acc 0.7196 top5_acc 0.9222 2021-06-06 00:04:13,447 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-06 00:04:13,507 - mmaction - INFO - mean_acc 0.7196 2021-06-06 00:04:13,508 - mmaction - INFO - Epoch(val) [22][38] top1_acc: 0.7196, top5_acc: 0.9222, mean_class_accuracy: 0.7196 2021-06-06 00:11:47,513 - mmaction - INFO - Epoch [23][20/38] lr: 1.500e-05, eta: 0:19:03, time: 22.696, data_time: 21.762, memory: 10388, top1_acc: 0.7766, top5_acc: 0.9510, loss_cls: 0.8496, loss: 0.8496, grad_norm: 2.6980 2021-06-06 00:12:06,019 - mmaction - INFO - Saving checkpoint at 23 epochs 2021-06-06 00:19:37,151 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-06 00:19:37,200 - mmaction - INFO - top1_acc 0.7183 top5_acc 0.9255 2021-06-06 00:19:37,200 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-06 00:19:37,265 - mmaction - INFO - mean_acc 0.7183 2021-06-06 00:19:37,266 - mmaction - INFO - Epoch(val) [23][38] top1_acc: 0.7183, top5_acc: 0.9255, mean_class_accuracy: 0.7183 2021-06-06 00:27:16,870 - mmaction - INFO - Epoch [24][20/38] lr: 1.500e-05, eta: 0:11:20, time: 22.978, data_time: 22.062, memory: 10388, top1_acc: 0.7568, top5_acc: 0.9276, loss_cls: 0.9246, loss: 0.9246, grad_norm: 2.7206 2021-06-06 00:27:35,524 - mmaction - INFO - Saving checkpoint at 24 epochs 2021-06-06 00:35:11,980 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-06 00:35:12,029 - mmaction - INFO - top1_acc 0.7176 top5_acc 0.9255 2021-06-06 00:35:12,029 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-06 00:35:12,094 - mmaction - INFO - mean_acc 0.7176 2021-06-06 00:35:12,095 - mmaction - INFO - Epoch(val) [24][38] top1_acc: 0.7176, top5_acc: 0.9255, mean_class_accuracy: 0.7176 2021-06-06 00:42:54,227 - mmaction - INFO - Epoch [25][20/38] lr: 1.500e-05, eta: 0:03:38, time: 23.104, data_time: 22.172, memory: 10388, top1_acc: 0.7661, top5_acc: 0.9385, loss_cls: 0.8723, loss: 0.8723, grad_norm: 2.6596 2021-06-06 00:43:12,589 - mmaction - INFO - Saving checkpoint at 25 epochs 2021-06-06 00:50:38,125 - mmaction - INFO - Evaluating top_k_accuracy ... 2021-06-06 00:50:38,169 - mmaction - INFO - top1_acc 0.7157 top5_acc 0.9275 2021-06-06 00:50:38,169 - mmaction - INFO - Evaluating mean_class_accuracy ... 2021-06-06 00:50:38,230 - mmaction - INFO - mean_acc 0.7157 2021-06-06 00:50:38,230 - mmaction - INFO - Epoch(val) [25][38] top1_acc: 0.7157, top5_acc: 0.9275, mean_class_accuracy: 0.7157