Closed FarranYang closed 2 years ago
我使用了自己的数据集,格式与ILSVRC一样,也都转为了cocovid格式。
[>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>> ] 5936/6186, 6.0 task/s, elapsed: 983s, ETA: 41s^CTraceback (most recent call last):
File "/data/yangjiahui/envs/torch15/lib/python3.8/runpy.py", line 194, in _run_module_as_main
return _run_code(code, main_globals, None,
File "/data/yangjiahui/envs/torch15/lib/python3.8/runpy.py", line 87, in _run_code
exec(code, run_globals)
File "/data/yangjiahui/envs/torch15/lib/python3.8/site-packages/torch/distributed/launch.py", line 263, in
The progress bar will be stuck since it only counts the processed images in rank 0, while the different ranks process different number of images in MMTracking.
When the progress bar is stuck, please use nvidia-smi
to observe whether there is a gpu still processing images. If so, please wait. If your dataset has a very long video, it is reasonable to be stuck for a long while.
thank you for your reply! when it stucks, I saw the processes by nviadia-smi,they keep running. I will wait for a longer period!
The progress bar will be stuck since it only counts the processed images in rank 0, while the different ranks process different number of images in MMTracking. When the progress bar is stuck, please use
nvidia-smi
to observe whether there is a gpu still processing images. If so, please wait. If your dataset has a very long video, it is reasonable to be stuck for a long while.
Hi,I have tried to run training process, but still got stuck in the evaluation . This time I waited for 21 hours and the process was not released, no errors were reported and no new logs were generated.
This is my stucked log: File "/data/yangjiahui/envs/torch15/lib/python3.8/site-packages/torch/nn/modules/module.py", line 550, in call result = self.forward(*input, *kwargs) File "/data/yangjiahui/envs/torch15/lib/python3.8/site-packages/mmcv/runner/fp16_utils.py", line 186, in new_func return old_func(args, **kwargs) File "/data/yangjiahui/VIDProject/mmtracking/mmtrack/models/roi_heads/roi_extractors/temporal_roi_align.py", line 195, in forward ref_roi_feats = self.most_similar_roi_align( File "/data/yangjiahui/VIDProject/mmtracking/mmtrack/models/roi_heads/roi_extractors/temporal_roi_align.py", line 175, in most_similar_roi_align ref_roi_feats = torch.cat((ref_roi_feats, one_ref_roi_feats), RuntimeError: CUDA out of memory. Tried to allocate 402.00 MiB (GPU 3; 10.76 GiB total capacity; 3.32 GiB already allocated; 152.56 MiB free; 3.95 GiB reserved in total by PyTorch) [>>>>>>>>>>>>>>>>>>>>>>> ] 5592/6186, 10.5 task/s, elapsed: 534s, ETA: 57s
The progress bar will be stuck since it only counts the processed images in rank 0, while the different ranks process different number of images in MMTracking. When the progress bar is stuck, please use
nvidia-smi
to observe whether there is a gpu still processing images. If so, please wait. If your dataset has a very long video, it is reasonable to be stuck for a long while.
我最长的一条视频,总帧数不到400,这能导致评估时间过长吗?
If your dataset has different categories with ImageNet VID, you need modify the CLASSES
in ImagenetVIDDataset
.
As you mentioned, I have changed the CLASSES in ImagenetVIDDataset and num_classes in base/models/faster_rcnn_r50_dc5.py,it still stucks.o(╥﹏╥)o
thank you for your reply, I rebuild my environment and it works well now!
我的日志如下:
2021-12-26 16:04:19,060 - mmtrack - INFO - Environment info:
sys.platform: linux Python: 3.8.12 (default, Oct 12 2021, 13:49:34) [GCC 7.5.0] CUDA available: True GPU 0,1,2: GeForce RTX 2080 Ti CUDA_HOME: /usr/local/cuda NVCC: Cuda compilation tools, release 10.0, V10.0.130 GCC: gcc (Ubuntu 5.4.0-6ubuntu1~16.04.12) 5.4.0 20160609 PyTorch: 1.5.0 PyTorch compiling details: PyTorch built with:
TorchVision: 0.6.0a0+82fd1c8 OpenCV: 4.5.4 MMCV: 1.4.1 MMCV Compiler: GCC 7.3 MMCV CUDA Compiler: 10.1 MMTracking: 0.8.0+
2021-12-26 16:04:19,061 - mmtrack - INFO - Distributed training: True 2021-12-26 16:04:19,761 - mmtrack - INFO - Config: model = dict( detector=dict( type='FasterRCNN', backbone=dict( type='ResNet', depth=50, num_stages=4, out_indices=(3, ), strides=(1, 2, 2, 1), dilations=(1, 1, 1, 2), frozen_stages=1, norm_cfg=dict(type='BN', requires_grad=True), norm_eval=True, style='pytorch'), neck=dict( type='ChannelMapper', in_channels=[2048], out_channels=512, kernel_size=3), rpn_head=dict( type='RPNHead', in_channels=512, feat_channels=512, anchor_generator=dict( type='AnchorGenerator', scales=[4, 8, 16, 32], ratios=[0.5, 1.0, 2.0], strides=[16]), bbox_coder=dict( type='DeltaXYWHBBoxCoder', target_means=[0.0, 0.0, 0.0, 0.0], target_stds=[1.0, 1.0, 1.0, 1.0]), loss_cls=dict( type='CrossEntropyLoss', use_sigmoid=True, loss_weight=1.0), loss_bbox=dict( type='SmoothL1Loss', beta=0.1111111111111111, loss_weight=1.0)), roi_head=dict( type='SelsaRoIHead', bbox_roi_extractor=dict( type='TemporalRoIAlign', roi_layer=dict( type='RoIAlign', output_size=7, sampling_ratio=2), out_channels=512, featmap_strides=[16], num_most_similar_points=2, num_temporal_attention_blocks=4), bbox_head=dict( type='SelsaBBoxHead', in_channels=512, fc_out_channels=1024, roi_feat_size=7, num_classes=30, bbox_coder=dict( type='DeltaXYWHBBoxCoder', target_means=[0.0, 0.0, 0.0, 0.0], target_stds=[0.2, 0.2, 0.2, 0.2]), reg_class_agnostic=False, loss_cls=dict( type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0), loss_bbox=dict(type='SmoothL1Loss', beta=1.0, loss_weight=1.0), num_shared_fcs=3, aggregator=dict( type='SelsaAggregator', in_channels=1024, num_attention_blocks=16))), train_cfg=dict( rpn=dict( assigner=dict( type='MaxIoUAssigner', pos_iou_thr=0.7, neg_iou_thr=0.3, min_pos_iou=0.3, ignore_iof_thr=-1), sampler=dict( type='RandomSampler', num=256, pos_fraction=0.5, neg_pos_ub=-1, add_gt_as_proposals=False), allowed_border=0, pos_weight=-1, debug=False), rpn_proposal=dict( nms_pre=6000, max_per_img=600, nms=dict(type='nms', iou_threshold=0.7), min_bbox_size=0), rcnn=dict( assigner=dict( type='MaxIoUAssigner', pos_iou_thr=0.5, neg_iou_thr=0.5, min_pos_iou=0.5, ignore_iof_thr=-1), sampler=dict( type='RandomSampler', num=256, pos_fraction=0.25, neg_pos_ub=-1, add_gt_as_proposals=True), pos_weight=-1, debug=False)), test_cfg=dict( rpn=dict( nms_pre=6000, max_per_img=300, nms=dict(type='nms', iou_threshold=0.7), min_bbox_size=0), rcnn=dict( score_thr=0.0001, nms=dict(type='nms', iou_threshold=0.5), max_per_img=100))), type='SELSA') dataset_type = 'ImagenetVIDDataset' data_root = 'data/FALD_VID/' img_norm_cfg = dict( mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True) train_pipeline = [ dict(type='LoadMultiImagesFromFile'), dict(type='SeqLoadAnnotations', with_bbox=True, with_track=True), dict(type='SeqResize', img_scale=(1000, 600), keep_ratio=True), dict(type='SeqRandomFlip', share_params=True, flip_ratio=0.5), dict( type='SeqNormalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='SeqPad', size_divisor=16), dict( type='VideoCollect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_instance_ids']), dict(type='ConcatVideoReferences'), dict(type='SeqDefaultFormatBundle', ref_prefix='ref') ] test_pipeline = [ dict(type='LoadMultiImagesFromFile'), dict(type='SeqResize', img_scale=(1000, 600), keep_ratio=True), dict(type='SeqRandomFlip', share_params=True, flip_ratio=0.0), dict( type='SeqNormalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='SeqPad', size_divisor=16), dict( type='VideoCollect', keys=['img'], meta_keys=('num_left_ref_imgs', 'frame_stride')), dict(type='ConcatVideoReferences'), dict(type='MultiImagesToTensor', ref_prefix='ref'), dict(type='ToList') ] data = dict( samples_per_gpu=1, workers_per_gpu=2, train=dict( type='ImagenetVIDDataset', ann_file= 'data/FALD_VID/COCOVIDannotations/imagenet_vid_train_every10frames.json', img_prefix='data/FALD_VID/Data/VID', ref_img_sampler=dict( num_ref_imgs=2, frame_range=9, filter_key_img=False, method='bilateral_uniform'), pipeline=[ dict(type='LoadMultiImagesFromFile'), dict(type='SeqLoadAnnotations', with_bbox=True, with_track=True), dict(type='SeqResize', img_scale=(1000, 600), keep_ratio=True), dict(type='SeqRandomFlip', share_params=True, flip_ratio=0.5), dict( type='SeqNormalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='SeqPad', size_divisor=16), dict( type='VideoCollect', keys=['img', 'gt_bboxes', 'gt_labels', 'gt_instance_ids']), dict(type='ConcatVideoReferences'), dict(type='SeqDefaultFormatBundle', ref_prefix='ref') ]), val=dict( type='ImagenetVIDDataset', ann_file='data/FALD_VID/annotations/imagenet_vid_val.json', img_prefix='data/FALD_VID/Data/VID', ref_img_sampler=dict( num_ref_imgs=14, frame_range=[-7, 7], method='test_with_adaptive_stride'), pipeline=[ dict(type='LoadMultiImagesFromFile'), dict(type='SeqResize', img_scale=(1000, 600), keep_ratio=True), dict(type='SeqRandomFlip', share_params=True, flip_ratio=0.0), dict( type='SeqNormalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='SeqPad', size_divisor=16), dict( type='VideoCollect', keys=['img'], meta_keys=('num_left_ref_imgs', 'frame_stride')), dict(type='ConcatVideoReferences'), dict(type='MultiImagesToTensor', ref_prefix='ref'), dict(type='ToList') ], test_mode=True), test=dict( type='ImagenetVIDDataset', ann_file='data/FALD_VID/annotations/imagenet_vid_val.json', img_prefix='data/FALD_VID/Data/VID', ref_img_sampler=dict( num_ref_imgs=14, frame_range=[-7, 7], method='test_with_adaptive_stride'), pipeline=[ dict(type='LoadMultiImagesFromFile'), dict(type='SeqResize', img_scale=(1000, 600), keep_ratio=True), dict(type='SeqRandomFlip', share_params=True, flip_ratio=0.0), dict( type='SeqNormalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='SeqPad', size_divisor=16), dict( type='VideoCollect', keys=['img'], meta_keys=('num_left_ref_imgs', 'frame_stride')), dict(type='ConcatVideoReferences'), dict(type='MultiImagesToTensor', ref_prefix='ref'), dict(type='ToList') ], test_mode=True)) optimizer = dict(type='SGD', lr=0.005, momentum=0.9, weight_decay=0.0001) optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2)) checkpoint_config = dict(interval=1) log_config = dict(interval=50, hooks=[dict(type='TextLoggerHook')]) dist_params = dict(backend='nccl') log_level = 'INFO' load_from = None resume_from = None workflow = [('train', 1)] lr_config = dict( policy='step', warmup='linear', warmup_iters=500, warmup_ratio=0.3333333333333333, step=[2, 5]) total_epochs = 4 evaluation = dict(metric=['bbox'], interval=4) work_dir = './work_dirs/20211226_001_try3/' gpu_ids = range(0, 1)
2021-12-26 16:04:24,438 - mmtrack - INFO - Set random seed to 2034425034, deterministic: False 2021-12-26 16:04:25,201 - mmtrack - INFO - initialize ResNet with init_cfg [{'type': 'Kaiming', 'layer': 'Conv2d'}, {'type': 'Constant', 'val': 1, 'layer': ['_BatchNorm', 'GroupNorm']}] 2021-12-26 16:04:25,466 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,467 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,468 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,470 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,471 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,472 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,473 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,475 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,477 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,479 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,481 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,482 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,484 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,490 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,496 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,500 - mmtrack - INFO - initialize Bottleneck with init_cfg {'type': 'Constant', 'val': 0, 'override': {'name': 'norm3'}} 2021-12-26 16:04:25,523 - mmtrack - INFO - initialize ChannelMapper with init_cfg {'type': 'Xavier', 'layer': 'Conv2d', 'distribution': 'uniform'} 2021-12-26 16:04:25,583 - mmtrack - INFO - initialize RPNHead with init_cfg {'type': 'Normal', 'layer': 'Conv2d', 'std': 0.01} 2021-12-26 16:04:25,637 - mmtrack - INFO - initialize SelsaBBoxHead with init_cfg [{'type': 'Normal', 'std': 0.01, 'override': {'name': 'fc_cls'}}, {'type': 'Normal', 'std': 0.001, 'override': {'name': 'fc_reg'}}, {'type': 'Xavier', 'distribution': 'uniform', 'override': [{'name': 'shared_fcs'}, {'name': 'cls_fcs'}, {'name': 'reg_fcs'}]}] Name of parameter - Initialization information
detector.backbone.conv1.weight - torch.Size([64, 3, 7, 7]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.bn1.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.bn1.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.conv1.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.0.bn1.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.bn1.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.conv2.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.0.bn2.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.bn2.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.conv3.weight - torch.Size([256, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.0.bn3.weight - torch.Size([256]): ConstantInit: val=0, bias=0
detector.backbone.layer1.0.bn3.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.downsample.0.weight - torch.Size([256, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.0.downsample.1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.0.downsample.1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.1.conv1.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.1.bn1.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.1.bn1.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.1.conv2.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.1.bn2.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.1.bn2.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.1.conv3.weight - torch.Size([256, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.1.bn3.weight - torch.Size([256]): ConstantInit: val=0, bias=0
detector.backbone.layer1.1.bn3.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.2.conv1.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.2.bn1.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.2.bn1.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.2.conv2.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.2.bn2.weight - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.2.bn2.bias - torch.Size([64]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer1.2.conv3.weight - torch.Size([256, 64, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer1.2.bn3.weight - torch.Size([256]): ConstantInit: val=0, bias=0
detector.backbone.layer1.2.bn3.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.conv1.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.0.bn1.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.bn1.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.conv2.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.0.bn2.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.bn2.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.conv3.weight - torch.Size([512, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.0.bn3.weight - torch.Size([512]): ConstantInit: val=0, bias=0
detector.backbone.layer2.0.bn3.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.downsample.0.weight - torch.Size([512, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.0.downsample.1.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.0.downsample.1.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.1.conv1.weight - torch.Size([128, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.1.bn1.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.1.bn1.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.1.conv2.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.1.bn2.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.1.bn2.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.1.conv3.weight - torch.Size([512, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.1.bn3.weight - torch.Size([512]): ConstantInit: val=0, bias=0
detector.backbone.layer2.1.bn3.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.2.conv1.weight - torch.Size([128, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.2.bn1.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.2.bn1.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.2.conv2.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.2.bn2.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.2.bn2.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.2.conv3.weight - torch.Size([512, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.2.bn3.weight - torch.Size([512]): ConstantInit: val=0, bias=0
detector.backbone.layer2.2.bn3.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.3.conv1.weight - torch.Size([128, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.3.bn1.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.3.bn1.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.3.conv2.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.3.bn2.weight - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.3.bn2.bias - torch.Size([128]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer2.3.conv3.weight - torch.Size([512, 128, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer2.3.bn3.weight - torch.Size([512]): ConstantInit: val=0, bias=0
detector.backbone.layer2.3.bn3.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.conv1.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.0.bn1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.bn1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.0.bn2.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.bn2.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.conv3.weight - torch.Size([1024, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.0.bn3.weight - torch.Size([1024]): ConstantInit: val=0, bias=0
detector.backbone.layer3.0.bn3.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.downsample.0.weight - torch.Size([1024, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.0.downsample.1.weight - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.0.downsample.1.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.1.conv1.weight - torch.Size([256, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.1.bn1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.1.bn1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.1.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.1.bn2.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.1.bn2.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.1.conv3.weight - torch.Size([1024, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.1.bn3.weight - torch.Size([1024]): ConstantInit: val=0, bias=0
detector.backbone.layer3.1.bn3.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.2.conv1.weight - torch.Size([256, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.2.bn1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.2.bn1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.2.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.2.bn2.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.2.bn2.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.2.conv3.weight - torch.Size([1024, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.2.bn3.weight - torch.Size([1024]): ConstantInit: val=0, bias=0
detector.backbone.layer3.2.bn3.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.3.conv1.weight - torch.Size([256, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.3.bn1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.3.bn1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.3.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.3.bn2.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.3.bn2.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.3.conv3.weight - torch.Size([1024, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.3.bn3.weight - torch.Size([1024]): ConstantInit: val=0, bias=0
detector.backbone.layer3.3.bn3.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.4.conv1.weight - torch.Size([256, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.4.bn1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.4.bn1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.4.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.4.bn2.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.4.bn2.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.4.conv3.weight - torch.Size([1024, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.4.bn3.weight - torch.Size([1024]): ConstantInit: val=0, bias=0
detector.backbone.layer3.4.bn3.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.5.conv1.weight - torch.Size([256, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.5.bn1.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.5.bn1.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.5.conv2.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.5.bn2.weight - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.5.bn2.bias - torch.Size([256]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer3.5.conv3.weight - torch.Size([1024, 256, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer3.5.bn3.weight - torch.Size([1024]): ConstantInit: val=0, bias=0
detector.backbone.layer3.5.bn3.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.conv1.weight - torch.Size([512, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.0.bn1.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.bn1.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.0.bn2.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.bn2.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.conv3.weight - torch.Size([2048, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.0.bn3.weight - torch.Size([2048]): ConstantInit: val=0, bias=0
detector.backbone.layer4.0.bn3.bias - torch.Size([2048]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.downsample.0.weight - torch.Size([2048, 1024, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.0.downsample.1.weight - torch.Size([2048]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.0.downsample.1.bias - torch.Size([2048]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.1.conv1.weight - torch.Size([512, 2048, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.1.bn1.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.1.bn1.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.1.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.1.bn2.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.1.bn2.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.1.conv3.weight - torch.Size([2048, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.1.bn3.weight - torch.Size([2048]): ConstantInit: val=0, bias=0
detector.backbone.layer4.1.bn3.bias - torch.Size([2048]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.2.conv1.weight - torch.Size([512, 2048, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.2.bn1.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.2.bn1.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.2.conv2.weight - torch.Size([512, 512, 3, 3]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.2.bn2.weight - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.2.bn2.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.backbone.layer4.2.conv3.weight - torch.Size([2048, 512, 1, 1]): KaimingInit: a=0, mode=fan_out, nonlinearity=relu, distribution =normal, bias=0
detector.backbone.layer4.2.bn3.weight - torch.Size([2048]): ConstantInit: val=0, bias=0
detector.backbone.layer4.2.bn3.bias - torch.Size([2048]): The value is the same before and after calling
init_weights
of SELSAdetector.neck.convs.0.conv.weight - torch.Size([512, 2048, 3, 3]): XavierInit: gain=1, distribution=uniform, bias=0
detector.neck.convs.0.conv.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.rpn_head.rpn_conv.weight - torch.Size([512, 512, 3, 3]): NormalInit: mean=0, std=0.01, bias=0
detector.rpn_head.rpn_conv.bias - torch.Size([512]): NormalInit: mean=0, std=0.01, bias=0
detector.rpn_head.rpn_cls.weight - torch.Size([12, 512, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
detector.rpn_head.rpn_cls.bias - torch.Size([12]): NormalInit: mean=0, std=0.01, bias=0
detector.rpn_head.rpn_reg.weight - torch.Size([48, 512, 1, 1]): NormalInit: mean=0, std=0.01, bias=0
detector.rpn_head.rpn_reg.bias - torch.Size([48]): NormalInit: mean=0, std=0.01, bias=0
detector.roi_head.bbox_roi_extractor.embed_network.conv.weight - torch.Size([512, 512, 3, 3]): Initialized by user-defined
init_weights
in ConvModuledetector.roi_head.bbox_roi_extractor.embed_network.conv.bias - torch.Size([512]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.fc_cls.weight - torch.Size([31, 1024]): NormalInit: mean=0, std=0.01, bias=0
detector.roi_head.bbox_head.fc_cls.bias - torch.Size([31]): NormalInit: mean=0, std=0.01, bias=0
detector.roi_head.bbox_head.fc_reg.weight - torch.Size([120, 1024]): NormalInit: mean=0, std=0.001, bias=0
detector.roi_head.bbox_head.fc_reg.bias - torch.Size([120]): NormalInit: mean=0, std=0.001, bias=0
detector.roi_head.bbox_head.shared_fcs.0.weight - torch.Size([1024, 25088]): XavierInit: gain=1, distribution=uniform, bias=0
detector.roi_head.bbox_head.shared_fcs.0.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
detector.roi_head.bbox_head.shared_fcs.1.weight - torch.Size([1024, 1024]): XavierInit: gain=1, distribution=uniform, bias=0
detector.roi_head.bbox_head.shared_fcs.1.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
detector.roi_head.bbox_head.shared_fcs.2.weight - torch.Size([1024, 1024]): XavierInit: gain=1, distribution=uniform, bias=0
detector.roi_head.bbox_head.shared_fcs.2.bias - torch.Size([1024]): XavierInit: gain=1, distribution=uniform, bias=0
detector.roi_head.bbox_head.aggregator.0.fc_embed.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.fc_embed.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.ref_fc_embed.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.ref_fc_embed.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.fc.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.fc.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.ref_fc.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.0.ref_fc.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.fc_embed.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.fc_embed.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.ref_fc_embed.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.ref_fc_embed.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.fc.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.fc.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.ref_fc.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.1.ref_fc.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.fc_embed.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.fc_embed.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.ref_fc_embed.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.ref_fc_embed.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.fc.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.fc.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.ref_fc.weight - torch.Size([1024, 1024]): The value is the same before and after calling
init_weights
of SELSAdetector.roi_head.bbox_head.aggregator.2.ref_fc.bias - torch.Size([1024]): The value is the same before and after calling
init_weights
of SELSA2021-12-26 16:04:28,460 - mmtrack - INFO - Start running, host: user-lbyjh@admin.cluster.local, work_dir: /data/yangjiahui/VIDProject/mmtracking/work_dirs/20211226_001_try3 2021-12-26 16:04:28,461 - mmtrack - INFO - Hooks will be executed in the following order: before_run: (VERY_HIGH ) StepLrUpdaterHook
(NORMAL ) CheckpointHook
(NORMAL ) DistEvalHook
(VERY_LOW ) TextLoggerHook
before_train_epoch: (VERY_HIGH ) StepLrUpdaterHook
(NORMAL ) DistSamplerSeedHook
(NORMAL ) DistEvalHook
(LOW ) IterTimerHook
(VERY_LOW ) TextLoggerHook
before_train_iter: (VERY_HIGH ) StepLrUpdaterHook
(NORMAL ) DistEvalHook
(LOW ) IterTimerHook
after_train_iter: (ABOVE_NORMAL) OptimizerHook
(NORMAL ) CheckpointHook
(NORMAL ) DistEvalHook
(LOW ) IterTimerHook
(VERY_LOW ) TextLoggerHook
after_train_epoch: (NORMAL ) CheckpointHook
(NORMAL ) DistEvalHook
(VERY_LOW ) TextLoggerHook
before_val_epoch: (NORMAL ) DistSamplerSeedHook
(LOW ) IterTimerHook
(VERY_LOW ) TextLoggerHook
before_val_iter: (LOW ) IterTimerHook
after_val_iter: (LOW ) IterTimerHook
after_val_epoch: (VERY_LOW ) TextLoggerHook
after_run: (VERY_LOW ) TextLoggerHook
2021-12-26 16:04:28,461 - mmtrack - INFO - workflow: [('train', 1)], max: 4 epochs 2021-12-26 16:04:28,461 - mmtrack - INFO - Checkpoints will be saved to /data/yangjiahui/VIDProject/mmtracking/work_dirs/20211226_001_try3 by HardDiskBackend. 2021-12-26 16:05:00,501 - mmtrack - INFO - Saving checkpoint at 1 epochs 2021-12-26 16:05:32,658 - mmtrack - INFO - Saving checkpoint at 2 epochs 2021-12-26 16:06:04,769 - mmtrack - INFO - Saving checkpoint at 3 epochs 2021-12-26 16:06:37,068 - mmtrack - INFO - Saving checkpoint at 4 epochs