search

open-mmlab / mmdetection

OpenMMLab Detection Toolbox and Benchmark
https://mmdetection.readthedocs.io
Apache License 2.0
29.33k stars 9.42k forks source link

[YOLOX] Is Prefetch Dataloader in your future plan? #5993

Closed RaymondByc closed 2 years ago

RaymondByc commented 3 years ago

We compare the training time of mmdet-based YOLOX and the official code. And our experiments show the mmdet code lag behind so much (2.5 days VS 4 days).

The main reason we found is the lack of the async prefetch mechanism in dataloader, which saves a lot of time in moving tensor from CPU to GPU.

Do you have a plan to add the mechanism?

hhaAndroid commented 3 years ago

@RaymondByc Thank you for your feedback, but there is no difference between the official code we ran and the mmdet code. Can you provide further information?

RaymondByc commented 3 years ago

The details can be found in data_prefetcher

hhaAndroid commented 3 years ago

@RaymondByc I mean do you have a detailed comparison log?

RaymondByc commented 3 years ago

The log of the official code.

2021-08-31 07:03:32.516 | INFO | yolox.core.trainer:before_train:126 - args: Namespace(batch_size=64, cache=False, ckpt=None, devices=8, dist_backend='nccl', dist_url=None, exp_file='exps/default/yolox_s.py', experiment_name='yolox_s', fp16=False, machine_rank=0, name=None, num_machines=1, occupy=False, opts=[], resume=False, start_epoch=None) 2021-08-31 07:03:32.520 | INFO | yolox.core.trainer:before_train:127 - exp value: ╒══════════════════╤════════════════════════════╕ │ keys │ values │ ╞══════════════════╪════════════════════════════╡ │ seed │ None │ ├──────────────────┼────────────────────────────┤ │ output_dir │ './YOLOX_outputs' │ ├──────────────────┼────────────────────────────┤ │ print_interval │ 10 │ ├──────────────────┼────────────────────────────┤ │ eval_interval │ 10 │ ├──────────────────┼────────────────────────────┤ │ num_classes │ 80 │ ├──────────────────┼────────────────────────────┤ │ depth │ 0.33 │ ├──────────────────┼────────────────────────────┤ │ width │ 0.5 │ ├──────────────────┼────────────────────────────┤ │ data_num_workers │ 4 │ ├──────────────────┼────────────────────────────┤ │ input_size │ (640, 640) │ ├──────────────────┼────────────────────────────┤ │ multiscale_range │ 5 │ ├──────────────────┼────────────────────────────┤ │ data_dir │ None │ ├──────────────────┼────────────────────────────┤ │ train_ann │ 'instances_train2017.json' │ ├──────────────────┼────────────────────────────┤ │ val_ann │ 'instances_val2017.json' │ ├──────────────────┼────────────────────────────┤ │ mosaic_prob │ 1.0 │ ├──────────────────┼────────────────────────────┤ │ mixup_prob │ 1.0 │ ├──────────────────┼────────────────────────────┤ │ hsv_prob │ 1.0 │ ├──────────────────┼────────────────────────────┤ │ flip_prob │ 0.5 │ ├──────────────────┼────────────────────────────┤ │ degrees │ 10.0 │ ├──────────────────┼────────────────────────────┤ │ translate │ 0.1 │ ├──────────────────┼────────────────────────────┤ │ mosaic_scale │ (0.1, 2) │ ├──────────────────┼────────────────────────────┤ │ mixup_scale │ (0.5, 1.5) │ ├──────────────────┼────────────────────────────┤ │ shear │ 2.0 │ ├──────────────────┼────────────────────────────┤ │ perspective │ 0.0 │ ├──────────────────┼────────────────────────────┤ │ enable_mixup │ True │ ├──────────────────┼────────────────────────────┤ │ warmup_epochs │ 5 │ ├──────────────────┼────────────────────────────┤ │ max_epoch │ 300 │ ├──────────────────┼────────────────────────────┤ │ warmup_lr │ 0 │ ├──────────────────┼────────────────────────────┤ │ basic_lr_per_img │ 0.00015625 │ ├──────────────────┼────────────────────────────┤ │ scheduler │ 'yoloxwarmcos' │ ├──────────────────┼────────────────────────────┤ │ no_aug_epochs │ 15 │ ├──────────────────┼────────────────────────────┤ │ min_lr_ratio │ 0.05 │ ├──────────────────┼────────────────────────────┤ │ ema │ True │ ├──────────────────┼────────────────────────────┤ │ weight_decay │ 0.0005 │ ├──────────────────┼────────────────────────────┤ │ momentum │ 0.9 │ ├──────────────────┼────────────────────────────┤ │ exp_name │ 'yolox_s' │ ├──────────────────┼────────────────────────────┤ │ test_size │ (640, 640) │ ├──────────────────┼────────────────────────────┤ │ test_conf │ 0.01 │ ├──────────────────┼────────────────────────────┤ │ nmsthre │ 0.65 │ ╘══════════════════╧════════════════════════════╛ 2021-08-31 07:03:40.361 | INFO | yolox.core.trainer:before_train:133 - Model Summary: Params: 8.97M, Gflops: 26.81 2021-08-31 07:03:40.410 | INFO | yolox.data.datasets.coco:init:45 - loading annotations into memory... 2021-08-31 07:03:56.685 | INFO | yolox.data.datasets.coco:init:45 - Done (t=16.27s) 2021-08-31 07:03:56.686 | INFO | pycocotools.coco:init:89 - creating index... 2021-08-31 07:04:00.331 | INFO | pycocotools.coco:init:89 - index created! 2021-08-31 07:04:33.172 | INFO | yolox.core.trainer:before_train:151 - init prefetcher, this might take one minute or less... 2021-08-31 07:07:59.045 | INFO | yolox.data.datasets.coco:init:45 - loading annotations into memory... 2021-08-31 07:07:59.479 | INFO | yolox.data.datasets.coco:init:45 - Done (t=0.43s) 2021-08-31 07:07:59.479 | INFO | pycocotools.coco:init:89 - creating index... 2021-08-31 07:07:59.526 | INFO | pycocotools.coco:init:89 - index created! 2021-08-31 07:08:00.875 | INFO | yolox.core.trainer:before_train:179 - Training start... 2021-08-31 07:08:00.878 | INFO | yolox.core.trainer:before_train:180 - DistributedDataParallel( (module): YOLOX( (backbone): YOLOPAFPN( (backbone): CSPDarknet( (stem): Focus( (conv): BaseConv( (conv): Conv2d(12, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(32, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (dark2): Sequential( (0): BaseConv( (conv): Conv2d(32, 64, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): CSPLayer( (conv1): BaseConv( (conv): Conv2d(64, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(32, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(64, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(32, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(32, 32, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(32, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(32, 32, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(32, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) ) (dark3): Sequential( (0): BaseConv( (conv): Conv2d(64, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): CSPLayer( (conv1): BaseConv( (conv): Conv2d(128, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(128, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (1): Bottleneck( (conv1): BaseConv( (conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (2): Bottleneck( (conv1): BaseConv( (conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) ) (dark4): Sequential( (0): BaseConv( (conv): Conv2d(128, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): CSPLayer( (conv1): BaseConv( (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (1): Bottleneck( (conv1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (2): Bottleneck( (conv1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) ) (dark5): Sequential( (0): BaseConv( (conv): Conv2d(256, 512, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn): BatchNorm2d(512, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): SPPBottleneck( (conv1): BaseConv( (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): ModuleList( (0): MaxPool2d(kernel_size=5, stride=1, padding=2, dilation=1, ceil_mode=False) (1): MaxPool2d(kernel_size=9, stride=1, padding=4, dilation=1, ceil_mode=False) (2): MaxPool2d(kernel_size=13, stride=1, padding=6, dilation=1, ceil_mode=False) ) (conv2): BaseConv( (conv): Conv2d(1024, 512, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(512, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (2): CSPLayer( (conv1): BaseConv( (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(512, 512, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(512, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) ) ) (upsample): Upsample(scale_factor=2.0, mode=nearest) (lateral_conv0): BaseConv( (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (C3_p4): CSPLayer( (conv1): BaseConv( (conv): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) (reduce_conv1): BaseConv( (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (C3_p3): CSPLayer( (conv1): BaseConv( (conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(256, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(64, 64, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(64, 64, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(64, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) (bu_conv2): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (C3_n3): CSPLayer( (conv1): BaseConv( (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) (bu_conv1): BaseConv( (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(2, 2), padding=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (C3_n4): CSPLayer( (conv1): BaseConv( (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(512, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv3): BaseConv( (conv): Conv2d(512, 512, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(512, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (m): Sequential( (0): Bottleneck( (conv1): BaseConv( (conv): Conv2d(256, 256, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (conv2): BaseConv( (conv): Conv2d(256, 256, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(256, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) ) ) (head): YOLOXHead( (cls_convs): ModuleList( (0): Sequential( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (1): Sequential( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (2): Sequential( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) (reg_convs): ModuleList( (0): Sequential( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (1): Sequential( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (2): Sequential( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(128, 128, kernel_size=(3, 3), stride=(1, 1), padding=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) ) (cls_preds): ModuleList( (0): Conv2d(128, 80, kernel_size=(1, 1), stride=(1, 1)) (1): Conv2d(128, 80, kernel_size=(1, 1), stride=(1, 1)) (2): Conv2d(128, 80, kernel_size=(1, 1), stride=(1, 1)) ) (reg_preds): ModuleList( (0): Conv2d(128, 4, kernel_size=(1, 1), stride=(1, 1)) (1): Conv2d(128, 4, kernel_size=(1, 1), stride=(1, 1)) (2): Conv2d(128, 4, kernel_size=(1, 1), stride=(1, 1)) ) (obj_preds): ModuleList( (0): Conv2d(128, 1, kernel_size=(1, 1), stride=(1, 1)) (1): Conv2d(128, 1, kernel_size=(1, 1), stride=(1, 1)) (2): Conv2d(128, 1, kernel_size=(1, 1), stride=(1, 1)) ) (stems): ModuleList( (0): BaseConv( (conv): Conv2d(128, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (1): BaseConv( (conv): Conv2d(256, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) (2): BaseConv( (conv): Conv2d(512, 128, kernel_size=(1, 1), stride=(1, 1), bias=False) (bn): BatchNorm2d(128, eps=0.001, momentum=0.03, affine=True, track_running_stats=True) (act): SiLU(inplace=True) ) ) (l1_loss): L1Loss() (bcewithlog_loss): BCEWithLogitsLoss() (iou_loss): IOUloss() ) ) ) 2021-08-31 07:08:00.878 | INFO | yolox.core.trainer:before_epoch:188 - ---> start train epoch1 2021-08-31 07:08:06.767 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 10/1849, mem: 6357Mb, iter_time: 0.588s, data_time: 0.001s, total_loss: 16.9, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 10.2, cls_loss: 2.0, lr: 1.170e-08, size: 640, ETA: 3 days, 18:38:06 2021-08-31 07:08:11.010 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 20/1849, mem: 6589Mb, iter_time: 0.423s, data_time: 0.002s, total_loss: 18.4, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 11.9, cls_loss: 1.8, lr: 4.680e-08, size: 672, ETA: 3 days, 5:56:30 2021-08-31 07:08:15.061 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 30/1849, mem: 6589Mb, iter_time: 0.402s, data_time: 0.007s, total_loss: 15.8, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 9.0, cls_loss: 2.1, lr: 1.053e-07, size: 576, ETA: 3 days, 0:36:46 2021-08-31 07:08:19.764 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 40/1849, mem: 6784Mb, iter_time: 0.469s, data_time: 0.002s, total_loss: 15.7, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.7, cls_loss: 2.3, lr: 1.872e-07, size: 704, ETA: 3 days, 0:31:37 2021-08-31 07:08:23.225 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 50/1849, mem: 6784Mb, iter_time: 0.345s, data_time: 0.002s, total_loss: 13.2, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 6.6, cls_loss: 1.9, lr: 2.925e-07, size: 512, ETA: 2 days, 20:38:29 2021-08-31 07:08:26.369 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 60/1849, mem: 6784Mb, iter_time: 0.310s, data_time: 0.002s, total_loss: 17.3, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 10.0, cls_loss: 2.5, lr: 4.212e-07, size: 672, ETA: 2 days, 17:09:29 2021-08-31 07:08:29.476 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 70/1849, mem: 6784Mb, iter_time: 0.309s, data_time: 0.002s, total_loss: 15.1, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 8.3, cls_loss: 2.0, lr: 5.733e-07, size: 672, ETA: 2 days, 14:39:33 2021-08-31 07:08:32.394 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 80/1849, mem: 6784Mb, iter_time: 0.290s, data_time: 0.002s, total_loss: 13.4, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 6.9, cls_loss: 1.8, lr: 7.488e-07, size: 512, ETA: 2 days, 12:24:13 2021-08-31 07:08:36.684 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 90/1849, mem: 6784Mb, iter_time: 0.427s, data_time: 0.002s, total_loss: 14.8, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.1, cls_loss: 2.0, lr: 9.477e-07, size: 608, ETA: 2 days, 12:59:52 2021-08-31 07:08:39.691 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 100/1849, mem: 6784Mb, iter_time: 0.299s, data_time: 0.002s, total_loss: 14.7, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 7.8, cls_loss: 2.1, lr: 1.170e-06, size: 512, ETA: 2 days, 11:30:15 2021-08-31 07:08:45.202 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 110/1849, mem: 7420Mb, iter_time: 0.549s, data_time: 0.004s, total_loss: 21.9, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 15.3, cls_loss: 1.8, lr: 1.416e-06, size: 800, ETA: 2 days, 13:47:01 2021-08-31 07:08:48.518 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 120/1849, mem: 7420Mb, iter_time: 0.330s, data_time: 0.002s, total_loss: 17.3, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 10.6, cls_loss: 1.9, lr: 1.685e-06, size: 672, ETA: 2 days, 12:52:20 2021-08-31 07:08:51.519 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 130/1849, mem: 7420Mb, iter_time: 0.299s, data_time: 0.002s, total_loss: 14.6, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 8.1, cls_loss: 1.8, lr: 1.977e-06, size: 512, ETA: 2 days, 11:43:39 2021-08-31 07:08:54.981 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 140/1849, mem: 7420Mb, iter_time: 0.345s, data_time: 0.001s, total_loss: 16.5, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.8, cls_loss: 2.1, lr: 2.293e-06, size: 640, ETA: 2 days, 11:15:35 2021-08-31 07:08:58.386 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 150/1849, mem: 7420Mb, iter_time: 0.340s, data_time: 0.001s, total_loss: 16.4, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.8, cls_loss: 1.9, lr: 2.633e-06, size: 640, ETA: 2 days, 10:47:44 2021-08-31 07:09:01.254 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 160/1849, mem: 7420Mb, iter_time: 0.286s, data_time: 0.002s, total_loss: 13.5, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 6.9, cls_loss: 1.9, lr: 2.995e-06, size: 480, ETA: 2 days, 9:52:25 2021-08-31 07:09:06.447 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 170/1849, mem: 7420Mb, iter_time: 0.518s, data_time: 0.002s, total_loss: 19.3, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 12.7, cls_loss: 1.8, lr: 3.381e-06, size: 768, ETA: 2 days, 11:09:42 2021-08-31 07:09:09.509 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 180/1849, mem: 7420Mb, iter_time: 0.305s, data_time: 0.002s, total_loss: 15.8, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 9.2, cls_loss: 1.9, lr: 3.791e-06, size: 672, ETA: 2 days, 10:29:16 2021-08-31 07:09:13.170 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 190/1849, mem: 7420Mb, iter_time: 0.365s, data_time: 0.003s, total_loss: 16.6, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.9, cls_loss: 2.0, lr: 4.224e-06, size: 800, ETA: 2 days, 10:21:49 2021-08-31 07:09:16.175 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 200/1849, mem: 7420Mb, iter_time: 0.298s, data_time: 0.002s, total_loss: 14.7, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.1, cls_loss: 2.0, lr: 4.680e-06, size: 512, ETA: 2 days, 9:44:30 2021-08-31 07:09:19.671 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 210/1849, mem: 7420Mb, iter_time: 0.347s, data_time: 0.001s, total_loss: 14.9, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 8.0, cls_loss: 2.3, lr: 5.160e-06, size: 640, ETA: 2 days, 9:32:15 2021-08-31 07:09:22.969 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 220/1849, mem: 7420Mb, iter_time: 0.329s, data_time: 0.002s, total_loss: 20.2, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 13.6, cls_loss: 1.9, lr: 5.663e-06, size: 768, ETA: 2 days, 9:13:24 2021-08-31 07:09:26.241 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 230/1849, mem: 7420Mb, iter_time: 0.326s, data_time: 0.001s, total_loss: 18.2, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 11.7, cls_loss: 1.7, lr: 6.189e-06, size: 640, ETA: 2 days, 8:55:14 2021-08-31 07:09:31.109 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 240/1849, mem: 7420Mb, iter_time: 0.485s, data_time: 0.002s, total_loss: 15.4, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.6, cls_loss: 2.1, lr: 6.739e-06, size: 736, ETA: 2 days, 9:39:26 2021-08-31 07:09:34.250 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 250/1849, mem: 7420Mb, iter_time: 0.313s, data_time: 0.002s, total_loss: 16.7, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 10.1, cls_loss: 1.9, lr: 7.313e-06, size: 672, ETA: 2 days, 9:16:43 2021-08-31 07:09:37.542 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 260/1849, mem: 7420Mb, iter_time: 0.328s, data_time: 0.002s, total_loss: 16.3, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.7, cls_loss: 1.8, lr: 7.909e-06, size: 704, ETA: 2 days, 9:01:06 2021-08-31 07:09:41.216 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 270/1849, mem: 7420Mb, iter_time: 0.366s, data_time: 0.002s, total_loss: 15.4, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.8, cls_loss: 1.9, lr: 8.529e-06, size: 800, ETA: 2 days, 8:59:38 2021-08-31 07:09:44.641 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 280/1849, mem: 7420Mb, iter_time: 0.342s, data_time: 0.001s, total_loss: 16.3, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.8, cls_loss: 1.8, lr: 9.173e-06, size: 640, ETA: 2 days, 8:50:08 2021-08-31 07:09:47.030 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 290/1849, mem: 7420Mb, iter_time: 0.238s, data_time: 0.002s, total_loss: 14.2, iou_loss: 4.8, l1_loss: 0.0, conf_loss: 7.7, cls_loss: 1.7, lr: 9.840e-06, size: 480, ETA: 2 days, 8:08:10 2021-08-31 07:09:50.558 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 300/1849, mem: 7420Mb, iter_time: 0.352s, data_time: 0.002s, total_loss: 18.7, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 12.0, cls_loss: 2.0, lr: 1.053e-05, size: 736, ETA: 2 days, 8:04:13 2021-08-31 07:09:53.699 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 310/1849, mem: 7420Mb, iter_time: 0.313s, data_time: 0.002s, total_loss: 14.8, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 8.1, cls_loss: 2.1, lr: 1.124e-05, size: 672, ETA: 2 days, 7:49:02 2021-08-31 07:09:56.555 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 320/1849, mem: 7420Mb, iter_time: 0.285s, data_time: 0.002s, total_loss: 13.7, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 7.0, cls_loss: 2.0, lr: 1.198e-05, size: 576, ETA: 2 days, 7:26:35 2021-08-31 07:09:59.529 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 330/1849, mem: 7420Mb, iter_time: 0.294s, data_time: 0.002s, total_loss: 13.4, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 6.8, cls_loss: 1.9, lr: 1.274e-05, size: 512, ETA: 2 days, 7:08:10 2021-08-31 07:10:03.079 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 340/1849, mem: 7420Mb, iter_time: 0.354s, data_time: 0.001s, total_loss: 14.9, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 8.2, cls_loss: 2.2, lr: 1.353e-05, size: 640, ETA: 2 days, 7:07:03 2021-08-31 07:10:06.145 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 350/1849, mem: 7420Mb, iter_time: 0.306s, data_time: 0.002s, total_loss: 14.7, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.2, cls_loss: 1.8, lr: 1.433e-05, size: 608, ETA: 2 days, 6:53:14 2021-08-31 07:10:09.268 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 360/1849, mem: 7420Mb, iter_time: 0.310s, data_time: 0.002s, total_loss: 15.0, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 8.5, cls_loss: 1.8, lr: 1.516e-05, size: 608, ETA: 2 days, 6:41:23 2021-08-31 07:10:12.631 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 370/1849, mem: 7420Mb, iter_time: 0.335s, data_time: 0.001s, total_loss: 14.5, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 7.6, cls_loss: 2.3, lr: 1.602e-05, size: 640, ETA: 2 days, 6:36:19 2021-08-31 07:10:16.569 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 380/1849, mem: 7420Mb, iter_time: 0.393s, data_time: 0.002s, total_loss: 12.9, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 6.3, cls_loss: 2.1, lr: 1.689e-05, size: 544, ETA: 2 days, 6:45:35 2021-08-31 07:10:19.722 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 390/1849, mem: 7420Mb, iter_time: 0.315s, data_time: 0.002s, total_loss: 13.6, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 7.1, cls_loss: 2.0, lr: 1.780e-05, size: 608, ETA: 2 days, 6:35:47 2021-08-31 07:10:22.230 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 400/1849, mem: 7420Mb, iter_time: 0.249s, data_time: 0.002s, total_loss: 12.9, iou_loss: 4.4, l1_loss: 0.0, conf_loss: 6.2, cls_loss: 2.2, lr: 1.872e-05, size: 480, ETA: 2 days, 6:11:16 2021-08-31 07:10:25.341 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 410/1849, mem: 7420Mb, iter_time: 0.310s, data_time: 0.002s, total_loss: 15.7, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.1, cls_loss: 1.9, lr: 1.967e-05, size: 672, ETA: 2 days, 6:01:47 2021-08-31 07:10:28.238 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 420/1849, mem: 7420Mb, iter_time: 0.287s, data_time: 0.002s, total_loss: 13.5, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 7.0, cls_loss: 1.9, lr: 2.064e-05, size: 512, ETA: 2 days, 5:47:46 2021-08-31 07:10:31.726 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 430/1849, mem: 7420Mb, iter_time: 0.347s, data_time: 0.001s, total_loss: 13.5, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 6.9, cls_loss: 2.1, lr: 2.163e-05, size: 640, ETA: 2 days, 5:47:09 2021-08-31 07:10:35.298 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 440/1849, mem: 7420Mb, iter_time: 0.356s, data_time: 0.002s, total_loss: 15.3, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 8.6, cls_loss: 2.1, lr: 2.265e-05, size: 768, ETA: 2 days, 5:48:27 2021-08-31 07:10:38.103 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 450/1849, mem: 7420Mb, iter_time: 0.279s, data_time: 0.003s, total_loss: 13.0, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 6.4, cls_loss: 1.9, lr: 2.369e-05, size: 544, ETA: 2 days, 5:33:59 2021-08-31 07:10:41.370 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 460/1849, mem: 7420Mb, iter_time: 0.326s, data_time: 0.002s, total_loss: 14.4, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 7.7, cls_loss: 2.0, lr: 2.476e-05, size: 704, ETA: 2 days, 5:29:30 2021-08-31 07:10:44.173 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 470/1849, mem: 7420Mb, iter_time: 0.278s, data_time: 0.003s, total_loss: 12.9, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 6.3, cls_loss: 2.0, lr: 2.585e-05, size: 544, ETA: 2 days, 5:15:50 2021-08-31 07:10:47.645 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 480/1849, mem: 7420Mb, iter_time: 0.346s, data_time: 0.002s, total_loss: 14.3, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 7.9, cls_loss: 2.0, lr: 2.696e-05, size: 736, ETA: 2 days, 5:15:52 2021-08-31 07:10:50.661 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 490/1849, mem: 7420Mb, iter_time: 0.301s, data_time: 0.002s, total_loss: 16.2, iou_loss: 4.7, l1_loss: 0.0, conf_loss: 9.8, cls_loss: 1.7, lr: 2.809e-05, size: 672, ETA: 2 days, 5:07:16 2021-08-31 07:10:53.468 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 500/1849, mem: 7420Mb, iter_time: 0.280s, data_time: 0.002s, total_loss: 14.4, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 7.9, cls_loss: 2.0, lr: 2.925e-05, size: 512, ETA: 2 days, 4:55:10 2021-08-31 07:10:56.719 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 510/1849, mem: 7420Mb, iter_time: 0.322s, data_time: 0.002s, total_loss: 14.0, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 7.3, cls_loss: 2.1, lr: 3.043e-05, size: 672, ETA: 2 days, 4:51:12 2021-08-31 07:10:59.142 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 520/1849, mem: 7420Mb, iter_time: 0.240s, data_time: 0.002s, total_loss: 13.0, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 6.5, cls_loss: 2.0, lr: 3.164e-05, size: 480, ETA: 2 days, 4:32:50 2021-08-31 07:11:02.528 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 530/1849, mem: 7420Mb, iter_time: 0.338s, data_time: 0.002s, total_loss: 14.7, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 8.3, cls_loss: 1.9, lr: 3.287e-05, size: 704, ETA: 2 days, 4:32:09 2021-08-31 07:11:05.487 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 540/1849, mem: 7420Mb, iter_time: 0.294s, data_time: 0.002s, total_loss: 13.3, iou_loss: 4.4, l1_loss: 0.0, conf_loss: 6.7, cls_loss: 2.2, lr: 3.412e-05, size: 576, ETA: 2 days, 4:24:04 2021-08-31 07:11:08.857 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 550/1849, mem: 7420Mb, iter_time: 0.335s, data_time: 0.001s, total_loss: 13.7, iou_loss: 4.4, l1_loss: 0.0, conf_loss: 7.0, cls_loss: 2.2, lr: 3.539e-05, size: 640, ETA: 2 days, 4:23:02 2021-08-31 07:11:11.582 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 560/1849, mem: 7420Mb, iter_time: 0.270s, data_time: 0.002s, total_loss: 12.9, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 6.5, cls_loss: 1.9, lr: 3.669e-05, size: 544, ETA: 2 days, 4:11:20 2021-08-31 07:11:15.259 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 570/1849, mem: 7420Mb, iter_time: 0.367s, data_time: 0.002s, total_loss: 14.1, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 7.6, cls_loss: 2.0, lr: 3.801e-05, size: 800, ETA: 2 days, 4:15:43 2021-08-31 07:11:18.634 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 580/1849, mem: 7420Mb, iter_time: 0.336s, data_time: 0.002s, total_loss: 14.3, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 7.8, cls_loss: 2.0, lr: 3.936e-05, size: 704, ETA: 2 days, 4:15:08 2021-08-31 07:11:21.738 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 590/1849, mem: 7420Mb, iter_time: 0.309s, data_time: 0.002s, total_loss: 12.9, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 6.4, cls_loss: 1.9, lr: 4.073e-05, size: 608, ETA: 2 days, 4:10:22 2021-08-31 07:11:25.120 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 600/1849, mem: 7420Mb, iter_time: 0.337s, data_time: 0.002s, total_loss: 14.1, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 7.6, cls_loss: 2.1, lr: 4.212e-05, size: 768, ETA: 2 days, 4:10:00 2021-08-31 07:11:28.011 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 610/1849, mem: 7420Mb, iter_time: 0.288s, data_time: 0.002s, total_loss: 13.1, iou_loss: 4.6, l1_loss: 0.0, conf_loss: 6.7, cls_loss: 1.8, lr: 4.354e-05, size: 576, ETA: 2 days, 4:02:16 2021-08-31 07:11:30.826 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 620/1849, mem: 7420Mb, iter_time: 0.279s, data_time: 0.002s, total_loss: 12.7, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 6.3, cls_loss: 1.9, lr: 4.497e-05, size: 544, ETA: 2 days, 3:53:28 2021-08-31 07:11:34.465 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 630/1849, mem: 7420Mb, iter_time: 0.363s, data_time: 0.002s, total_loss: 14.8, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 8.4, cls_loss: 1.8, lr: 4.644e-05, size: 800, ETA: 2 days, 3:57:12 2021-08-31 07:11:37.721 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 640/1849, mem: 7420Mb, iter_time: 0.324s, data_time: 0.001s, total_loss: 12.4, iou_loss: 4.3, l1_loss: 0.0, conf_loss: 5.8, cls_loss: 2.2, lr: 4.792e-05, size: 640, ETA: 2 days, 3:55:15 2021-08-31 07:11:40.671 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 650/1849, mem: 7420Mb, iter_time: 0.294s, data_time: 0.002s, total_loss: 12.5, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 6.1, cls_loss: 2.0, lr: 4.943e-05, size: 512, ETA: 2 days, 3:49:01 2021-08-31 07:11:44.370 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 660/1849, mem: 7420Mb, iter_time: 0.368s, data_time: 0.002s, total_loss: 13.3, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 6.8, cls_loss: 2.3, lr: 5.097e-05, size: 800, ETA: 2 days, 3:53:22 2021-08-31 07:11:47.717 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 670/1849, mem: 7420Mb, iter_time: 0.333s, data_time: 0.002s, total_loss: 13.2, iou_loss: 4.3, l1_loss: 0.0, conf_loss: 6.7, cls_loss: 2.2, lr: 5.252e-05, size: 736, ETA: 2 days, 3:52:44 2021-08-31 07:11:51.389 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 680/1849, mem: 7420Mb, iter_time: 0.366s, data_time: 0.002s, total_loss: 13.7, iou_loss: 4.5, l1_loss: 0.0, conf_loss: 7.3, cls_loss: 1.9, lr: 5.410e-05, size: 800, ETA: 2 days, 3:56:36 2021-08-31 07:11:54.232 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 690/1849, mem: 7420Mb, iter_time: 0.283s, data_time: 0.002s, total_loss: 12.4, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 5.9, cls_loss: 2.3, lr: 5.570e-05, size: 544, ETA: 2 days, 3:49:19 2021-08-31 07:11:57.020 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 700/1849, mem: 7420Mb, iter_time: 0.277s, data_time: 0.002s, total_loss: 12.6, iou_loss: 4.4, l1_loss: 0.0, conf_loss: 6.1, cls_loss: 2.1, lr: 5.733e-05, size: 544, ETA: 2 days, 3:41:26 2021-08-31 07:12:00.332 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 710/1849, mem: 7420Mb, iter_time: 0.330s, data_time: 0.002s, total_loss: 13.1, iou_loss: 4.4, l1_loss: 0.0, conf_loss: 6.7, cls_loss: 2.0, lr: 5.898e-05, size: 704, ETA: 2 days, 3:40:34 2021-08-31 07:12:03.137 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 720/1849, mem: 7420Mb, iter_time: 0.279s, data_time: 0.003s, total_loss: 12.9, iou_loss: 4.4, l1_loss: 0.0, conf_loss: 6.3, cls_loss: 2.1, lr: 6.065e-05, size: 544, ETA: 2 days, 3:33:15 2021-08-31 07:12:06.042 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 730/1849, mem: 7420Mb, iter_time: 0.290s, data_time: 0.002s, total_loss: 11.7, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.4, lr: 6.235e-05, size: 544, ETA: 2 days, 3:27:27 2021-08-31 07:12:09.146 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 740/1849, mem: 7420Mb, iter_time: 0.308s, data_time: 0.002s, total_loss: 12.6, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 6.1, cls_loss: 2.4, lr: 6.407e-05, size: 672, ETA: 2 days, 3:24:08 2021-08-31 07:12:12.069 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 750/1849, mem: 7420Mb, iter_time: 0.291s, data_time: 0.002s, total_loss: 12.3, iou_loss: 4.3, l1_loss: 0.0, conf_loss: 5.9, cls_loss: 2.1, lr: 6.581e-05, size: 576, ETA: 2 days, 3:18:50 2021-08-31 07:12:15.184 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 760/1849, mem: 7420Mb, iter_time: 0.310s, data_time: 0.002s, total_loss: 12.6, iou_loss: 4.3, l1_loss: 0.0, conf_loss: 6.0, cls_loss: 2.2, lr: 6.758e-05, size: 608, ETA: 2 days, 3:15:52 2021-08-31 07:12:18.335 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 770/1849, mem: 7420Mb, iter_time: 0.313s, data_time: 0.002s, total_loss: 12.2, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 5.8, cls_loss: 2.2, lr: 6.937e-05, size: 672, ETA: 2 days, 3:13:26 2021-08-31 07:12:21.269 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 780/1849, mem: 7420Mb, iter_time: 0.293s, data_time: 0.002s, total_loss: 11.8, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.2, lr: 7.118e-05, size: 512, ETA: 2 days, 3:08:36 2021-08-31 07:12:24.552 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 790/1849, mem: 7420Mb, iter_time: 0.328s, data_time: 0.002s, total_loss: 12.3, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 5.8, cls_loss: 2.3, lr: 7.302e-05, size: 704, ETA: 2 days, 3:07:59 2021-08-31 07:12:27.281 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 800/1849, mem: 7420Mb, iter_time: 0.272s, data_time: 0.002s, total_loss: 11.4, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.3, lr: 7.488e-05, size: 544, ETA: 2 days, 3:00:55 2021-08-31 07:12:30.666 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 810/1849, mem: 7420Mb, iter_time: 0.338s, data_time: 0.002s, total_loss: 12.3, iou_loss: 4.3, l1_loss: 0.0, conf_loss: 5.9, cls_loss: 2.1, lr: 7.676e-05, size: 704, ETA: 2 days, 3:01:34 2021-08-31 07:12:34.151 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 820/1849, mem: 7420Mb, iter_time: 0.347s, data_time: 0.002s, total_loss: 12.4, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.7, cls_loss: 2.7, lr: 7.867e-05, size: 736, ETA: 2 days, 3:03:17 2021-08-31 07:12:37.532 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 830/1849, mem: 7420Mb, iter_time: 0.337s, data_time: 0.002s, total_loss: 11.9, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.5, cls_loss: 2.3, lr: 8.060e-05, size: 704, ETA: 2 days, 3:03:50 2021-08-31 07:12:40.950 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 840/1849, mem: 7420Mb, iter_time: 0.340s, data_time: 0.002s, total_loss: 12.3, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.8, cls_loss: 2.4, lr: 8.256e-05, size: 768, ETA: 2 days, 3:04:41 2021-08-31 07:12:43.422 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 850/1849, mem: 7420Mb, iter_time: 0.246s, data_time: 0.002s, total_loss: 11.9, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.5, cls_loss: 2.5, lr: 8.453e-05, size: 480, ETA: 2 days, 2:55:20 2021-08-31 07:12:46.858 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 860/1849, mem: 7420Mb, iter_time: 0.343s, data_time: 0.002s, total_loss: 12.1, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.5, cls_loss: 2.5, lr: 8.653e-05, size: 704, ETA: 2 days, 2:56:31 2021-08-31 07:12:50.256 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 870/1849, mem: 7420Mb, iter_time: 0.339s, data_time: 0.002s, total_loss: 11.8, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.3, lr: 8.856e-05, size: 704, ETA: 2 days, 2:57:18 2021-08-31 07:12:53.887 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 880/1849, mem: 7420Mb, iter_time: 0.362s, data_time: 0.004s, total_loss: 12.5, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 6.1, cls_loss: 2.3, lr: 9.060e-05, size: 800, ETA: 2 days, 3:00:28 2021-08-31 07:12:56.816 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 890/1849, mem: 7420Mb, iter_time: 0.292s, data_time: 0.003s, total_loss: 12.1, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 2.5, lr: 9.268e-05, size: 576, ETA: 2 days, 2:56:16 2021-08-31 07:13:00.281 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 900/1849, mem: 7420Mb, iter_time: 0.346s, data_time: 0.001s, total_loss: 12.4, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 6.0, cls_loss: 2.3, lr: 9.477e-05, size: 640, ETA: 2 days, 2:57:43 2021-08-31 07:13:03.008 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 910/1849, mem: 7420Mb, iter_time: 0.272s, data_time: 0.002s, total_loss: 11.8, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.5, lr: 9.689e-05, size: 544, ETA: 2 days, 2:51:36 2021-08-31 07:13:06.446 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 920/1849, mem: 7420Mb, iter_time: 0.341s, data_time: 0.001s, total_loss: 12.1, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 2.5, lr: 9.903e-05, size: 640, ETA: 2 days, 2:52:35 2021-08-31 07:13:09.773 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 930/1849, mem: 7420Mb, iter_time: 0.331s, data_time: 0.001s, total_loss: 11.5, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.3, lr: 1.012e-04, size: 640, ETA: 2 days, 2:52:31 2021-08-31 07:13:12.736 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 940/1849, mem: 7420Mb, iter_time: 0.294s, data_time: 0.002s, total_loss: 12.1, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.7, cls_loss: 2.4, lr: 1.034e-04, size: 576, ETA: 2 days, 2:48:53 2021-08-31 07:13:16.220 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 950/1849, mem: 7420Mb, iter_time: 0.347s, data_time: 0.002s, total_loss: 12.6, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 6.1, cls_loss: 2.4, lr: 1.056e-04, size: 704, ETA: 2 days, 2:50:24 2021-08-31 07:13:19.707 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 960/1849, mem: 7420Mb, iter_time: 0.347s, data_time: 0.002s, total_loss: 12.1, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 2.5, lr: 1.078e-04, size: 768, ETA: 2 days, 2:51:57 2021-08-31 07:13:23.202 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 970/1849, mem: 7420Mb, iter_time: 0.349s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.4, lr: 1.101e-04, size: 768, ETA: 2 days, 2:53:35 2021-08-31 07:13:26.713 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 980/1849, mem: 7420Mb, iter_time: 0.350s, data_time: 0.002s, total_loss: 13.1, iou_loss: 4.2, l1_loss: 0.0, conf_loss: 6.7, cls_loss: 2.2, lr: 1.124e-04, size: 768, ETA: 2 days, 2:55:19 2021-08-31 07:13:29.580 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 990/1849, mem: 7420Mb, iter_time: 0.286s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.6, lr: 1.147e-04, size: 544, ETA: 2 days, 2:51:02 2021-08-31 07:13:32.556 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1000/1849, mem: 7420Mb, iter_time: 0.297s, data_time: 0.002s, total_loss: 11.7, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.4, lr: 1.170e-04, size: 576, ETA: 2 days, 2:47:52 2021-08-31 07:13:35.386 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1010/1849, mem: 7420Mb, iter_time: 0.281s, data_time: 0.003s, total_loss: 11.1, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 4.7, cls_loss: 2.6, lr: 1.194e-04, size: 544, ETA: 2 days, 2:43:20 2021-08-31 07:13:38.850 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1020/1849, mem: 7420Mb, iter_time: 0.345s, data_time: 0.002s, total_loss: 12.1, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.7, cls_loss: 2.5, lr: 1.217e-04, size: 736, ETA: 2 days, 2:44:42 2021-08-31 07:13:42.248 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1030/1849, mem: 7420Mb, iter_time: 0.339s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.5, lr: 1.241e-04, size: 736, ETA: 2 days, 2:45:27 2021-08-31 07:13:45.861 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1040/1849, mem: 7420Mb, iter_time: 0.360s, data_time: 0.002s, total_loss: 12.8, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 6.3, cls_loss: 2.5, lr: 1.265e-04, size: 800, ETA: 2 days, 2:48:04 2021-08-31 07:13:49.223 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1050/1849, mem: 7420Mb, iter_time: 0.335s, data_time: 0.001s, total_loss: 11.5, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.6, lr: 1.290e-04, size: 640, ETA: 2 days, 2:48:27 2021-08-31 07:13:52.389 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1060/1849, mem: 7420Mb, iter_time: 0.315s, data_time: 0.002s, total_loss: 11.8, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.6, lr: 1.315e-04, size: 672, ETA: 2 days, 2:47:05 2021-08-31 07:13:55.563 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1070/1849, mem: 7420Mb, iter_time: 0.316s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.6, lr: 1.340e-04, size: 672, ETA: 2 days, 2:45:48 2021-08-31 07:13:58.785 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1080/1849, mem: 7420Mb, iter_time: 0.320s, data_time: 0.002s, total_loss: 11.8, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.6, lr: 1.365e-04, size: 672, ETA: 2 days, 2:44:54 2021-08-31 07:14:02.329 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1090/1849, mem: 7420Mb, iter_time: 0.353s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.6, lr: 1.390e-04, size: 800, ETA: 2 days, 2:46:46 2021-08-31 07:14:05.264 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1100/1849, mem: 7420Mb, iter_time: 0.292s, data_time: 0.003s, total_loss: 11.3, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.4, lr: 1.416e-04, size: 512, ETA: 2 days, 2:43:31 2021-08-31 07:14:08.728 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1110/1849, mem: 7420Mb, iter_time: 0.346s, data_time: 0.002s, total_loss: 11.9, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 2.4, lr: 1.442e-04, size: 768, ETA: 2 days, 2:44:46 2021-08-31 07:14:11.842 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1120/1849, mem: 7420Mb, iter_time: 0.310s, data_time: 0.002s, total_loss: 12.0, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.5, cls_loss: 2.7, lr: 1.468e-04, size: 608, ETA: 2 days, 2:43:05 2021-08-31 07:14:15.249 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1130/1849, mem: 7420Mb, iter_time: 0.339s, data_time: 0.002s, total_loss: 11.8, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 2.6, lr: 1.494e-04, size: 768, ETA: 2 days, 2:43:46 2021-08-31 07:14:18.266 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1140/1849, mem: 7420Mb, iter_time: 0.300s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.6, lr: 1.521e-04, size: 512, ETA: 2 days, 2:41:20 2021-08-31 07:14:21.583 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1150/1849, mem: 7420Mb, iter_time: 0.331s, data_time: 0.002s, total_loss: 11.8, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.6, lr: 1.547e-04, size: 704, ETA: 2 days, 2:41:22 2021-08-31 07:14:24.810 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1160/1849, mem: 7420Mb, iter_time: 0.321s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.6, lr: 1.574e-04, size: 672, ETA: 2 days, 2:40:37 2021-08-31 07:14:27.815 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1170/1849, mem: 7420Mb, iter_time: 0.300s, data_time: 0.002s, total_loss: 11.4, iou_loss: 4.1, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.3, lr: 1.602e-04, size: 512, ETA: 2 days, 2:38:13 2021-08-31 07:14:30.905 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1180/1849, mem: 7420Mb, iter_time: 0.308s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.8, lr: 1.629e-04, size: 576, ETA: 2 days, 2:36:30 2021-08-31 07:14:33.820 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1190/1849, mem: 7420Mb, iter_time: 0.291s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.4, lr: 1.657e-04, size: 576, ETA: 2 days, 2:33:27 2021-08-31 07:14:36.760 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1200/1849, mem: 7420Mb, iter_time: 0.291s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.5, lr: 1.685e-04, size: 576, ETA: 2 days, 2:30:32 2021-08-31 07:14:40.420 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1210/1849, mem: 7420Mb, iter_time: 0.364s, data_time: 0.003s, total_loss: 11.7, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.6, lr: 1.713e-04, size: 800, ETA: 2 days, 2:33:10 2021-08-31 07:14:43.543 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1220/1849, mem: 7420Mb, iter_time: 0.311s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.5, lr: 1.741e-04, size: 608, ETA: 2 days, 2:31:46 2021-08-31 07:14:46.681 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1230/1849, mem: 7420Mb, iter_time: 0.312s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.7, lr: 1.770e-04, size: 608, ETA: 2 days, 2:30:29 2021-08-31 07:14:50.364 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1240/1849, mem: 7420Mb, iter_time: 0.366s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.5, cls_loss: 2.5, lr: 1.799e-04, size: 800, ETA: 2 days, 2:33:13 2021-08-31 07:14:53.306 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1250/1849, mem: 7420Mb, iter_time: 0.293s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.4, lr: 1.828e-04, size: 512, ETA: 2 days, 2:30:33 2021-08-31 07:14:56.908 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1260/1849, mem: 7420Mb, iter_time: 0.359s, data_time: 0.002s, total_loss: 11.9, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 2.6, lr: 1.857e-04, size: 800, ETA: 2 days, 2:32:44 2021-08-31 07:15:00.433 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1270/1849, mem: 7420Mb, iter_time: 0.352s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.8, lr: 1.887e-04, size: 736, ETA: 2 days, 2:34:20 2021-08-31 07:15:03.336 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1280/1849, mem: 7420Mb, iter_time: 0.289s, data_time: 0.003s, total_loss: 11.6, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.8, lr: 1.917e-04, size: 544, ETA: 2 days, 2:31:22 2021-08-31 07:15:06.375 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1290/1849, mem: 7420Mb, iter_time: 0.303s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.4, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.8, lr: 1.947e-04, size: 608, ETA: 2 days, 2:29:27 2021-08-31 07:15:09.994 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1300/1849, mem: 7420Mb, iter_time: 0.360s, data_time: 0.002s, total_loss: 12.0, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.8, cls_loss: 2.8, lr: 1.977e-04, size: 800, ETA: 2 days, 2:31:37 2021-08-31 07:15:13.381 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1310/1849, mem: 7420Mb, iter_time: 0.338s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.7, lr: 2.008e-04, size: 704, ETA: 2 days, 2:32:12 2021-08-31 07:15:16.334 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1320/1849, mem: 7420Mb, iter_time: 0.294s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.9, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.5, lr: 2.039e-04, size: 576, ETA: 2 days, 2:29:44 2021-08-31 07:15:19.854 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1330/1849, mem: 7420Mb, iter_time: 0.350s, data_time: 0.001s, total_loss: 11.5, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.7, lr: 2.070e-04, size: 640, ETA: 2 days, 2:31:10 2021-08-31 07:15:22.923 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1340/1849, mem: 7420Mb, iter_time: 0.306s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.4, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.6, lr: 2.101e-04, size: 672, ETA: 2 days, 2:29:32 2021-08-31 07:15:26.446 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1350/1849, mem: 7420Mb, iter_time: 0.351s, data_time: 0.002s, total_loss: 12.1, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.7, cls_loss: 2.7, lr: 2.132e-04, size: 736, ETA: 2 days, 2:31:02 2021-08-31 07:15:29.416 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1360/1849, mem: 7420Mb, iter_time: 0.296s, data_time: 0.002s, total_loss: 11.4, iou_loss: 4.0, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.4, lr: 2.164e-04, size: 512, ETA: 2 days, 2:28:46 2021-08-31 07:15:32.240 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1370/1849, mem: 7420Mb, iter_time: 0.280s, data_time: 0.003s, total_loss: 11.2, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.5, lr: 2.196e-04, size: 544, ETA: 2 days, 2:25:29 2021-08-31 07:15:35.216 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1380/1849, mem: 7420Mb, iter_time: 0.297s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.7, lr: 2.228e-04, size: 512, ETA: 2 days, 2:23:21 2021-08-31 07:15:37.619 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1390/1849, mem: 7420Mb, iter_time: 0.239s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.8, lr: 2.261e-04, size: 480, ETA: 2 days, 2:17:25 2021-08-31 07:15:40.815 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1400/1849, mem: 7420Mb, iter_time: 0.318s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.7, lr: 2.293e-04, size: 608, ETA: 2 days, 2:16:47 2021-08-31 07:15:44.011 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1410/1849, mem: 7420Mb, iter_time: 0.319s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.6, lr: 2.326e-04, size: 608, ETA: 2 days, 2:16:10 2021-08-31 07:15:46.535 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1420/1849, mem: 7420Mb, iter_time: 0.251s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.8, lr: 2.359e-04, size: 480, ETA: 2 days, 2:11:12 2021-08-31 07:15:49.475 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1430/1849, mem: 7420Mb, iter_time: 0.292s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.8, lr: 2.393e-04, size: 576, ETA: 2 days, 2:08:55 2021-08-31 07:15:52.683 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1440/1849, mem: 7420Mb, iter_time: 0.319s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.5, lr: 2.426e-04, size: 672, ETA: 2 days, 2:08:22 2021-08-31 07:15:55.689 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1450/1849, mem: 7420Mb, iter_time: 0.299s, data_time: 0.002s, total_loss: 11.0, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 4.8, cls_loss: 2.6, lr: 2.460e-04, size: 576, ETA: 2 days, 2:06:36 2021-08-31 07:15:58.711 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1460/1849, mem: 7420Mb, iter_time: 0.301s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.4, lr: 2.494e-04, size: 608, ETA: 2 days, 2:04:57 2021-08-31 07:16:01.678 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1470/1849, mem: 7420Mb, iter_time: 0.294s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.9, lr: 2.528e-04, size: 608, ETA: 2 days, 2:02:53 2021-08-31 07:16:04.872 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1480/1849, mem: 7420Mb, iter_time: 0.317s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.8, lr: 2.563e-04, size: 672, ETA: 2 days, 2:02:20 2021-08-31 07:16:07.747 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1490/1849, mem: 7420Mb, iter_time: 0.287s, data_time: 0.002s, total_loss: 11.1, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.4, lr: 2.598e-04, size: 544, ETA: 2 days, 1:59:52 2021-08-31 07:16:11.139 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1500/1849, mem: 7420Mb, iter_time: 0.338s, data_time: 0.002s, total_loss: 12.2, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.6, cls_loss: 3.1, lr: 2.633e-04, size: 704, ETA: 2 days, 2:00:35 2021-08-31 07:16:14.396 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1510/1849, mem: 7420Mb, iter_time: 0.325s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.8, lr: 2.668e-04, size: 672, ETA: 2 days, 2:00:30 2021-08-31 07:16:17.446 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1520/1849, mem: 7420Mb, iter_time: 0.304s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.5, lr: 2.703e-04, size: 608, ETA: 2 days, 1:59:07 2021-08-31 07:16:20.780 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1530/1849, mem: 7420Mb, iter_time: 0.333s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.7, lr: 2.739e-04, size: 704, ETA: 2 days, 1:59:30 2021-08-31 07:16:24.200 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1540/1849, mem: 7420Mb, iter_time: 0.340s, data_time: 0.001s, total_loss: 11.3, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.7, lr: 2.775e-04, size: 640, ETA: 2 days, 2:00:21 2021-08-31 07:16:27.174 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1550/1849, mem: 7420Mb, iter_time: 0.297s, data_time: 0.002s, total_loss: 11.2, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.7, lr: 2.811e-04, size: 512, ETA: 2 days, 1:58:34 2021-08-31 07:16:30.281 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1560/1849, mem: 7420Mb, iter_time: 0.310s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.6, lr: 2.847e-04, size: 672, ETA: 2 days, 1:57:37 2021-08-31 07:16:33.205 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1570/1849, mem: 7420Mb, iter_time: 0.291s, data_time: 0.002s, total_loss: 12.0, iou_loss: 3.8, l1_loss: 0.0, conf_loss: 5.7, cls_loss: 2.6, lr: 2.884e-04, size: 576, ETA: 2 days, 1:55:34 2021-08-31 07:16:36.022 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1580/1849, mem: 7420Mb, iter_time: 0.280s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.8, lr: 2.921e-04, size: 544, ETA: 2 days, 1:52:54 2021-08-31 07:16:38.992 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1590/1849, mem: 7420Mb, iter_time: 0.296s, data_time: 0.002s, total_loss: 11.2, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 4.8, cls_loss: 2.8, lr: 2.958e-04, size: 512, ETA: 2 days, 1:51:10 2021-08-31 07:16:41.908 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1600/1849, mem: 7420Mb, iter_time: 0.290s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.8, lr: 2.995e-04, size: 512, ETA: 2 days, 1:49:09 2021-08-31 07:16:44.937 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1610/1849, mem: 7420Mb, iter_time: 0.301s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.7, lr: 3.033e-04, size: 672, ETA: 2 days, 1:47:46 2021-08-31 07:16:48.499 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1620/1849, mem: 7420Mb, iter_time: 0.354s, data_time: 0.002s, total_loss: 11.6, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.6, lr: 3.071e-04, size: 800, ETA: 2 days, 1:49:26 2021-08-31 07:16:51.191 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1630/1849, mem: 7420Mb, iter_time: 0.268s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.8, lr: 3.109e-04, size: 544, ETA: 2 days, 1:46:12 2021-08-31 07:16:53.993 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1640/1849, mem: 7420Mb, iter_time: 0.279s, data_time: 0.002s, total_loss: 11.2, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.5, lr: 3.147e-04, size: 544, ETA: 2 days, 1:43:38 2021-08-31 07:16:57.433 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1650/1849, mem: 7420Mb, iter_time: 0.343s, data_time: 0.001s, total_loss: 11.1, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.5, lr: 3.185e-04, size: 640, ETA: 2 days, 1:44:40 2021-08-31 07:17:00.642 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1660/1849, mem: 7420Mb, iter_time: 0.318s, data_time: 0.002s, total_loss: 11.5, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.9, lr: 3.224e-04, size: 672, ETA: 2 days, 1:44:18 2021-08-31 07:17:03.960 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1670/1849, mem: 7420Mb, iter_time: 0.331s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.7, lr: 3.263e-04, size: 736, ETA: 2 days, 1:44:37 2021-08-31 07:17:07.085 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1680/1849, mem: 7420Mb, iter_time: 0.311s, data_time: 0.002s, total_loss: 11.2, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.9, lr: 3.302e-04, size: 608, ETA: 2 days, 1:43:53 2021-08-31 07:17:10.557 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1690/1849, mem: 7420Mb, iter_time: 0.346s, data_time: 0.002s, total_loss: 12.3, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.9, cls_loss: 2.9, lr: 3.342e-04, size: 768, ETA: 2 days, 1:45:03 2021-08-31 07:17:13.004 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1700/1849, mem: 7420Mb, iter_time: 0.243s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.7, lr: 3.381e-04, size: 480, ETA: 2 days, 1:40:38 2021-08-31 07:17:16.408 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1710/1849, mem: 7420Mb, iter_time: 0.339s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.3, cls_loss: 2.6, lr: 3.421e-04, size: 736, ETA: 2 days, 1:41:26 2021-08-31 07:17:19.209 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1720/1849, mem: 7420Mb, iter_time: 0.279s, data_time: 0.002s, total_loss: 11.2, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.4, lr: 3.461e-04, size: 544, ETA: 2 days, 1:39:00 2021-08-31 07:17:22.649 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1730/1849, mem: 7420Mb, iter_time: 0.342s, data_time: 0.001s, total_loss: 11.6, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.8, lr: 3.502e-04, size: 640, ETA: 2 days, 1:39:57 2021-08-31 07:17:25.508 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1740/1849, mem: 7420Mb, iter_time: 0.285s, data_time: 0.002s, total_loss: 11.1, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 4.8, cls_loss: 2.8, lr: 3.542e-04, size: 544, ETA: 2 days, 1:37:52 2021-08-31 07:17:28.340 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1750/1849, mem: 7420Mb, iter_time: 0.283s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.8, lr: 3.583e-04, size: 544, ETA: 2 days, 1:35:40 2021-08-31 07:17:31.220 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1760/1849, mem: 7420Mb, iter_time: 0.286s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.3, l1_loss: 0.0, conf_loss: 5.2, cls_loss: 2.9, lr: 3.624e-04, size: 544, ETA: 2 days, 1:33:41 2021-08-31 07:17:34.799 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1770/1849, mem: 7420Mb, iter_time: 0.357s, data_time: 0.002s, total_loss: 11.7, iou_loss: 3.4, l1_loss: 0.0, conf_loss: 5.4, cls_loss: 2.9, lr: 3.665e-04, size: 736, ETA: 2 days, 1:35:25 2021-08-31 07:17:38.336 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1780/1849, mem: 7420Mb, iter_time: 0.353s, data_time: 0.001s, total_loss: 11.2, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 5.1, cls_loss: 2.6, lr: 3.707e-04, size: 640, ETA: 2 days, 1:36:55 2021-08-31 07:17:41.802 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1790/1849, mem: 7420Mb, iter_time: 0.346s, data_time: 0.002s, total_loss: 12.3, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 6.3, cls_loss: 2.5, lr: 3.749e-04, size: 736, ETA: 2 days, 1:38:02 2021-08-31 07:17:44.246 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1800/1849, mem: 7420Mb, iter_time: 0.243s, data_time: 0.002s, total_loss: 11.4, iou_loss: 3.4, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 3.1, lr: 3.791e-04, size: 480, ETA: 2 days, 1:33:54 2021-08-31 07:17:47.311 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1810/1849, mem: 7420Mb, iter_time: 0.305s, data_time: 0.003s, total_loss: 11.0, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 4.8, cls_loss: 2.5, lr: 3.833e-04, size: 512, ETA: 2 days, 1:32:57 2021-08-31 07:17:50.082 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1820/1849, mem: 7420Mb, iter_time: 0.275s, data_time: 0.002s, total_loss: 11.3, iou_loss: 3.6, l1_loss: 0.0, conf_loss: 5.0, cls_loss: 2.7, lr: 3.876e-04, size: 544, ETA: 2 days, 1:30:30 2021-08-31 07:17:53.081 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1830/1849, mem: 7420Mb, iter_time: 0.299s, data_time: 0.002s, total_loss: 11.2, iou_loss: 3.5, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.8, lr: 3.918e-04, size: 608, ETA: 2 days, 1:29:16 2021-08-31 07:17:56.440 | INFO | yolox.core.trainer:after_iter:246 - epoch: 1/300, iter: 1840/1849, mem: 7420Mb, iter_time: 0.334s, data_time: 0.028s, total_loss: 11.2, iou_loss: 3.7, l1_loss: 0.0, conf_loss: 4.9, cls_loss: 2.6, lr: 3.961e-04, size: 576, ETA: 2 days, 1:29:48 2021-08-31 07:17:59.268 | INFO | yolox.core.trainer:save_ckpt:318 - Save weights to ./YOLOX_outputs/yolox_s

RaymondByc commented 3 years ago

The log of MMdet is as followed:

2021-09-01 02:32:50,706 - mmdet - INFO - Environment info:

sys.platform: linux Python: 3.7.11 (default, Jul 27 2021, 14:32:16) [GCC 7.5.0] CUDA available: True GPU 0,1,2,3,4,5,6,7: Tesla V100-PCIE-32GB CUDA_HOME: /usr/local/cuda NVCC: Cuda compilation tools, release 10.1, V10.1.243 GCC: gcc (Ubuntu 7.5.0-3ubuntu1~18.04) 7.5.0 PyTorch: 1.8.1 PyTorch compiling details: PyTorch built with:

TorchVision: 0.9.1 OpenCV: 4.2.0 MMCV: 1.3.12 MMCV Compiler: GCC 7.5 MMCV CUDA Compiler: 10.1 MMDetection: 2.15.1+

2021-09-01 02:32:51,673 - mmdet - INFO - Distributed training: True 2021-09-01 02:32:52,567 - mmdet - INFO - Config: optimizer = dict( type='SGD', lr=0.01, momentum=0.9, weight_decay=0.0005, nesterov=True, paramwise_cfg=dict(norm_decay_mult=0.0, bias_decay_mult=0.0)) optimizer_config = dict(grad_clip=None) lr_config = dict( policy='YOLOX', warmup='exp', by_epoch=False, warmup_by_epoch=True, warmup_ratio=1, warmup_iters=5, num_last_epochs=15, min_lr_ratio=0.05) runner = dict(type='EpochBasedRunner', max_epochs=300) checkpoint_config = dict(interval=10) log_config = dict(interval=50, hooks=[dict(type='TextLoggerHook')]) custom_hooks = [ dict(type='YOLOXModeSwitchHook', num_last_epochs=15, priority=48), dict( type='SyncRandomSizeHook', ratio_range=(14, 26), img_scale=(640, 640), interval=10, priority=48), dict(type='SyncNormHook', num_last_epochs=15, interval=10, priority=48), dict(type='ExpMomentumEMAHook', resume_from=None, priority=49) ] dist_params = dict(backend='nccl') log_level = 'INFO' load_from = None resume_from = None workflow = [('train', 1)] model = dict( type='YOLOX', backbone=dict(type='CSPDarknet', deepen_factor=0.33, widen_factor=0.5), neck=dict( type='YOLOXPAFPN', in_channels=[128, 256, 512], out_channels=128, num_csp_blocks=1), bbox_head=dict( type='YOLOXHead', num_classes=80, in_channels=128, feat_channels=128), train_cfg=dict(assigner=dict(type='SimOTAAssigner', center_radius=2.5)), test_cfg=dict(score_thr=0.01, nms=dict(type='nms', iou_threshold=0.65))) dataset_type = 'CocoDataset' data_root = '/home/CAMERA_SHARE/data/coco/' img_norm_cfg = dict( mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True) img_scale = (640, 640) train_pipeline = [ dict(type='Mosaic', img_scale=(640, 640), pad_val=114.0), dict( type='RandomAffine', scaling_ratio_range=(0.1, 2), border=(-320, -320)), dict( type='MixUp', img_scale=(640, 640), ratio_range=(0.8, 1.6), pad_val=114.0), dict( type='PhotoMetricDistortion', brightness_delta=32, contrast_range=(0.5, 1.5), saturation_range=(0.5, 1.5), hue_delta=18), dict(type='RandomFlip', flip_ratio=0.5), dict(type='Resize', keep_ratio=True), dict(type='Pad', pad_to_square=True, pad_val=114.0), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']) ] train_dataset = dict( type='MultiImageMixDataset', dataset=dict( type='CocoDataset', ann_file= '/home/CAMERA_SHARE/data/coco/annotations/instances_train2017.json', img_prefix='/home/CAMERA_SHARE/data/coco/train2017/', pipeline=[ dict(type='LoadImageFromFile', to_float32=True), dict(type='LoadAnnotations', with_bbox=True) ], filter_empty_gt=False), pipeline=[ dict(type='Mosaic', img_scale=(640, 640), pad_val=114.0), dict( type='RandomAffine', scaling_ratio_range=(0.1, 2), border=(-320, -320)), dict( type='MixUp', img_scale=(640, 640), ratio_range=(0.8, 1.6), pad_val=114.0), dict( type='PhotoMetricDistortion', brightness_delta=32, contrast_range=(0.5, 1.5), saturation_range=(0.5, 1.5), hue_delta=18), dict(type='RandomFlip', flip_ratio=0.5), dict(type='Resize', keep_ratio=True), dict(type='Pad', pad_to_square=True, pad_val=114.0), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']) ], dynamic_scale=(640, 640)) test_pipeline = [ dict(type='LoadImageFromFile'), dict( type='MultiScaleFlipAug', img_scale=(640, 640), flip=False, transforms=[ dict(type='Resize', keep_ratio=True), dict(type='RandomFlip'), dict(type='Pad', size=(640, 640), pad_val=114.0), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img']) ]) ] data = dict( samples_per_gpu=8, workers_per_gpu=8, train=dict( type='MultiImageMixDataset', dataset=dict( type='CocoDataset', ann_file= '/home/CAMERA_SHARE/data/coco/annotations/instances_train2017.json', img_prefix='/home/CAMERA_SHARE/data/coco/train2017/', pipeline=[ dict(type='LoadImageFromFile', to_float32=True), dict(type='LoadAnnotations', with_bbox=True) ], filter_empty_gt=False), pipeline=[ dict(type='Mosaic', img_scale=(640, 640), pad_val=114.0), dict( type='RandomAffine', scaling_ratio_range=(0.1, 2), border=(-320, -320)), dict( type='MixUp', img_scale=(640, 640), ratio_range=(0.8, 1.6), pad_val=114.0), dict( type='PhotoMetricDistortion', brightness_delta=32, contrast_range=(0.5, 1.5), saturation_range=(0.5, 1.5), hue_delta=18), dict(type='RandomFlip', flip_ratio=0.5), dict(type='Resize', keep_ratio=True), dict(type='Pad', pad_to_square=True, pad_val=114.0), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img', 'gt_bboxes', 'gt_labels']) ], dynamic_scale=(640, 640)), val=dict( type='CocoDataset', ann_file= '/home/CAMERA_SHARE/data/coco/annotations/instances_val2017.json', img_prefix='/home/CAMERA_SHARE/data/coco/val2017/', pipeline=[ dict(type='LoadImageFromFile'), dict( type='MultiScaleFlipAug', img_scale=(640, 640), flip=False, transforms=[ dict(type='Resize', keep_ratio=True), dict(type='RandomFlip'), dict(type='Pad', size=(640, 640), pad_val=114.0), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img']) ]) ]), test=dict( type='CocoDataset', ann_file= '/home/CAMERA_SHARE/data/coco/annotations/instances_val2017.json', img_prefix='/home/CAMERA_SHARE/data/coco/val2017/', pipeline=[ dict(type='LoadImageFromFile'), dict( type='MultiScaleFlipAug', img_scale=(640, 640), flip=False, transforms=[ dict(type='Resize', keep_ratio=True), dict(type='RandomFlip'), dict(type='Pad', size=(640, 640), pad_val=114.0), dict( type='Normalize', mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True), dict(type='DefaultFormatBundle'), dict(type='Collect', keys=['img']) ]) ])) interval = 10 evaluation = dict(interval=10, metric='bbox') work_dir = './work_dirs/yolox_s_8x8_300e_coco_worker4' gpu_ids = range(0, 8)

2021-09-01 02:32:52,812 - mmdet - INFO - initialize CSPDarknet with init_cfg {'type': 'Kaiming', 'layer': 'Conv2d', 'a': 2.23606797749979, 'distribution': 'uniform', 'mode': 'fan_in', 'nonlinearity': 'leaky_relu'} 2021-09-01 02:32:52,857 - mmdet - INFO - initialize YOLOXPAFPN with init_cfg {'type': 'Kaiming', 'layer': 'Conv2d', 'a': 2.23606797749979, 'distribution': 'uniform', 'mode': 'fan_in', 'nonlinearity': 'leaky_relu'} 2021-09-01 02:32:52,886 - mmdet - INFO - initialize YOLOXHead with init_cfg {'type': 'Kaiming', 'layer': 'Conv2d', 'a': 2.23606797749979, 'distribution': 'uniform', 'mode': 'fan_in', 'nonlinearity': 'leaky_relu'} Name of parameter - Initialization information

backbone.stem.conv.conv.weight - torch.Size([32, 12, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stem.conv.bn.weight - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stem.conv.bn.bias - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.0.conv.weight - torch.Size([64, 32, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage1.0.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.0.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.main_conv.conv.weight - torch.Size([32, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage1.1.main_conv.bn.weight - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.main_conv.bn.bias - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.short_conv.conv.weight - torch.Size([32, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage1.1.short_conv.bn.weight - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.short_conv.bn.bias - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.final_conv.conv.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage1.1.final_conv.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.final_conv.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.blocks.0.conv1.conv.weight - torch.Size([32, 32, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage1.1.blocks.0.conv1.bn.weight - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.blocks.0.conv1.bn.bias - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.blocks.0.conv2.conv.weight - torch.Size([32, 32, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage1.1.blocks.0.conv2.bn.weight - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage1.1.blocks.0.conv2.bn.bias - torch.Size([32]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.0.conv.weight - torch.Size([128, 64, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.main_conv.conv.weight - torch.Size([64, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.main_conv.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.main_conv.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.short_conv.conv.weight - torch.Size([64, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.short_conv.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.short_conv.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.final_conv.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.final_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.final_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.0.conv1.conv.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.blocks.0.conv1.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.0.conv1.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.0.conv2.conv.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.blocks.0.conv2.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.0.conv2.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.1.conv1.conv.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.blocks.1.conv1.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.1.conv1.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.1.conv2.conv.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.blocks.1.conv2.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.1.conv2.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.2.conv1.conv.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.blocks.2.conv1.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.2.conv1.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.2.conv2.conv.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage2.1.blocks.2.conv2.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage2.1.blocks.2.conv2.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.0.conv.weight - torch.Size([256, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.0.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.0.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.main_conv.conv.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.main_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.main_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.short_conv.conv.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.short_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.short_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.final_conv.conv.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.final_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.final_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.0.conv1.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.blocks.0.conv1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.0.conv1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.0.conv2.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.blocks.0.conv2.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.0.conv2.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.1.conv1.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.blocks.1.conv1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.1.conv1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.1.conv2.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.blocks.1.conv2.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.1.conv2.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.2.conv1.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.blocks.2.conv1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.2.conv1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.2.conv2.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage3.1.blocks.2.conv2.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage3.1.blocks.2.conv2.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.0.conv.weight - torch.Size([512, 256, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.0.bn.weight - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.0.bn.bias - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.1.conv1.conv.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.1.conv1.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.1.conv1.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.1.conv2.conv.weight - torch.Size([512, 1024, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.1.conv2.bn.weight - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.1.conv2.bn.bias - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.main_conv.conv.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.2.main_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.main_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.short_conv.conv.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.2.short_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.short_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.final_conv.conv.weight - torch.Size([512, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.2.final_conv.bn.weight - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.final_conv.bn.bias - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.blocks.0.conv1.conv.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.2.blocks.0.conv1.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.blocks.0.conv1.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.blocks.0.conv2.conv.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

backbone.stage4.2.blocks.0.conv2.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

backbone.stage4.2.blocks.0.conv2.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.reduce_layers.0.conv.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.reduce_layers.0.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.reduce_layers.0.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.reduce_layers.1.conv.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.reduce_layers.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.reduce_layers.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.main_conv.conv.weight - torch.Size([128, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.0.main_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.main_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.short_conv.conv.weight - torch.Size([128, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.0.short_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.short_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.final_conv.conv.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.0.final_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.final_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.blocks.0.conv1.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.0.blocks.0.conv1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.blocks.0.conv1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.blocks.0.conv2.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.0.blocks.0.conv2.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.0.blocks.0.conv2.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.main_conv.conv.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.1.main_conv.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.main_conv.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.short_conv.conv.weight - torch.Size([64, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.1.short_conv.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.short_conv.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.final_conv.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.1.final_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.final_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.blocks.0.conv1.conv.weight - torch.Size([64, 64, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.1.blocks.0.conv1.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.blocks.0.conv1.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.blocks.0.conv2.conv.weight - torch.Size([64, 64, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.top_down_blocks.1.blocks.0.conv2.bn.weight - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.top_down_blocks.1.blocks.0.conv2.bn.bias - torch.Size([64]): The value is the same before and after calling init_weights of YOLOX

neck.downsamples.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.downsamples.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.downsamples.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.downsamples.1.conv.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.downsamples.1.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.downsamples.1.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.main_conv.conv.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.0.main_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.main_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.short_conv.conv.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.0.short_conv.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.short_conv.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.final_conv.conv.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.0.final_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.final_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.blocks.0.conv1.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.0.blocks.0.conv1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.blocks.0.conv1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.blocks.0.conv2.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.0.blocks.0.conv2.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.0.blocks.0.conv2.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.main_conv.conv.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.1.main_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.main_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.short_conv.conv.weight - torch.Size([256, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.1.short_conv.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.short_conv.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.final_conv.conv.weight - torch.Size([512, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.1.final_conv.bn.weight - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.final_conv.bn.bias - torch.Size([512]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.blocks.0.conv1.conv.weight - torch.Size([256, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.1.blocks.0.conv1.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.blocks.0.conv1.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.blocks.0.conv2.conv.weight - torch.Size([256, 256, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.bottom_up_blocks.1.blocks.0.conv2.bn.weight - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.bottom_up_blocks.1.blocks.0.conv2.bn.bias - torch.Size([256]): The value is the same before and after calling init_weights of YOLOX

neck.out_convs.0.conv.weight - torch.Size([128, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.out_convs.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.out_convs.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.out_convs.1.conv.weight - torch.Size([128, 256, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.out_convs.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.out_convs.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.out_convs.2.conv.weight - torch.Size([128, 512, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

neck.out_convs.2.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

neck.out_convs.2.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.0.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_cls_convs.0.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.0.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.0.1.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_cls_convs.0.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.0.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.1.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_cls_convs.1.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.1.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.1.1.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_cls_convs.1.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.1.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.2.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_cls_convs.2.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.2.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.2.1.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_cls_convs.2.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_cls_convs.2.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.0.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_reg_convs.0.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.0.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.0.1.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_reg_convs.0.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.0.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.1.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_reg_convs.1.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.1.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.1.1.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_reg_convs.1.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.1.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.2.0.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_reg_convs.2.0.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.2.0.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.2.1.conv.weight - torch.Size([128, 128, 3, 3]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_reg_convs.2.1.bn.weight - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_reg_convs.2.1.bn.bias - torch.Size([128]): The value is the same before and after calling init_weights of YOLOX

bbox_head.multi_level_conv_cls.0.weight - torch.Size([80, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_cls.0.bias - torch.Size([80]): Initialized by user-defined init_weights in YOLOXHead

bbox_head.multi_level_conv_cls.1.weight - torch.Size([80, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_cls.1.bias - torch.Size([80]): Initialized by user-defined init_weights in YOLOXHead

bbox_head.multi_level_conv_cls.2.weight - torch.Size([80, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_cls.2.bias - torch.Size([80]): Initialized by user-defined init_weights in YOLOXHead

bbox_head.multi_level_conv_reg.0.weight - torch.Size([4, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_reg.0.bias - torch.Size([4]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_reg.1.weight - torch.Size([4, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_reg.1.bias - torch.Size([4]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_reg.2.weight - torch.Size([4, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_reg.2.bias - torch.Size([4]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_obj.0.weight - torch.Size([1, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_obj.0.bias - torch.Size([1]): Initialized by user-defined init_weights in YOLOXHead

bbox_head.multi_level_conv_obj.1.weight - torch.Size([1, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_obj.1.bias - torch.Size([1]): Initialized by user-defined init_weights in YOLOXHead

bbox_head.multi_level_conv_obj.2.weight - torch.Size([1, 128, 1, 1]): KaimingInit: a=2.23606797749979, mode=fan_in, nonlinearity=leaky_relu, distribution =uniform, bias=0

bbox_head.multi_level_conv_obj.2.bias - torch.Size([1]): Initialized by user-defined init_weights in YOLOXHead
2021-09-01 02:33:25,535 - mmdet - INFO - Start running, host: b00586770@camera, work_dir: /home/b00586770/PycharmProjects/mmdetection-master/work_dirs/yolox_s_8x8_300e_coco_worker4 2021-09-01 02:33:25,535 - mmdet - INFO - Hooks will be executed in the following order: before_run: (VERY_HIGH ) YOLOXLrUpdaterHook
(49 ) ExpMomentumEMAHook
(NORMAL ) CheckpointHook
(NORMAL ) DistEvalHook
(VERY_LOW ) TextLoggerHook

before_train_epoch: (VERY_HIGH ) YOLOXLrUpdaterHook
(48 ) YOLOXModeSwitchHook
(48 ) SyncNormHook
(49 ) ExpMomentumEMAHook
(NORMAL ) DistSamplerSeedHook
(NORMAL ) DistEvalHook
(LOW ) IterTimerHook
(VERY_LOW ) TextLoggerHook

before_train_iter: (VERY_HIGH ) YOLOXLrUpdaterHook
(NORMAL ) DistEvalHook
(LOW ) IterTimerHook

after_train_iter: (ABOVE_NORMAL) OptimizerHook
(48 ) SyncRandomSizeHook
(49 ) ExpMomentumEMAHook
(NORMAL ) CheckpointHook
(NORMAL ) DistEvalHook
(LOW ) IterTimerHook
(VERY_LOW ) TextLoggerHook

after_train_epoch: (48 ) SyncNormHook
(49 ) ExpMomentumEMAHook
(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

2021-09-01 02:33:25,536 - mmdet - INFO - workflow: [('train', 1)], max: 300 epochs 2021-09-01 02:37:18,857 - mmdet - INFO - Epoch [1][50/1849] lr: 2.925e-07, eta: 29 days, 22:43:53, time: 4.665, data_time: 4.206, memory: 4721, loss_cls: 1.9623, loss_bbox: 4.7615, loss_obj: 9.2917, loss: 16.0154 2021-09-01 02:37:42,152 - mmdet - INFO - Epoch [1][100/1849] lr: 1.170e-06, eta: 16 days, 11:12:31, time: 0.466, data_time: 0.028, memory: 4731, loss_cls: 1.9756, loss_bbox: 4.7623, loss_obj: 9.2551, loss: 15.9930 2021-09-01 02:38:06,653 - mmdet - INFO - Epoch [1][150/1849] lr: 2.633e-06, eta: 12 days, 0:37:48, time: 0.490, data_time: 0.025, memory: 4731, loss_cls: 1.9596, loss_bbox: 4.7618, loss_obj: 9.2430, loss: 15.9644 2021-09-01 02:38:31,743 - mmdet - INFO - Epoch [1][200/1849] lr: 4.680e-06, eta: 9 days, 19:46:22, time: 0.502, data_time: 0.025, memory: 4762, loss_cls: 1.9206, loss_bbox: 4.7621, loss_obj: 9.1768, loss: 15.8595 2021-09-01 02:38:56,800 - mmdet - INFO - Epoch [1][250/1849] lr: 7.313e-06, eta: 8 days, 12:01:24, time: 0.501, data_time: 0.026, memory: 4798, loss_cls: 1.9228, loss_bbox: 4.7553, loss_obj: 9.0237, loss: 15.7018 2021-09-01 02:39:21,846 - mmdet - INFO - Epoch [1][300/1849] lr: 1.053e-05, eta: 7 days, 14:52:13, time: 0.501, data_time: 0.028, memory: 4798, loss_cls: 1.8791, loss_bbox: 4.7505, loss_obj: 8.8209, loss: 15.4504 2021-09-01 02:39:47,578 - mmdet - INFO - Epoch [1][350/1849] lr: 1.433e-05, eta: 7 days, 0:03:22, time: 0.515, data_time: 0.026, memory: 4798, loss_cls: 1.8480, loss_bbox: 4.7404, loss_obj: 8.5463, loss: 15.1348 2021-09-01 02:40:13,066 - mmdet - INFO - Epoch [1][400/1849] lr: 1.872e-05, eta: 6 days, 12:50:11, time: 0.509, data_time: 0.026, memory: 4798, loss_cls: 1.8588, loss_bbox: 4.7173, loss_obj: 8.2847, loss: 14.8607 2021-09-01 02:40:38,998 - mmdet - INFO - Epoch [1][450/1849] lr: 2.369e-05, eta: 6 days, 4:16:43, time: 0.519, data_time: 0.027, memory: 4798, loss_cls: 1.8752, loss_bbox: 4.6898, loss_obj: 8.0608, loss: 14.6258 2021-09-01 02:41:04,993 - mmdet - INFO - Epoch [1][500/1849] lr: 2.925e-05, eta: 5 days, 21:26:09, time: 0.519, data_time: 0.027, memory: 4798, loss_cls: 1.9003, loss_bbox: 4.6485, loss_obj: 7.7144, loss: 14.2633 2021-09-01 02:41:28,028 - mmdet - INFO - Epoch [1][550/1849] lr: 3.539e-05, eta: 5 days, 15:00:15, time: 0.460, data_time: 0.024, memory: 4798, loss_cls: 1.9559, loss_bbox: 4.5971, loss_obj: 7.4192, loss: 13.9721 2021-09-01 02:41:52,356 - mmdet - INFO - Epoch [1][600/1849] lr: 4.212e-05, eta: 5 days, 9:59:42, time: 0.487, data_time: 0.027, memory: 4798, loss_cls: 2.0289, loss_bbox: 4.5263, loss_obj: 7.0028, loss: 13.5579 2021-09-01 02:42:16,984 - mmdet - INFO - Epoch [1][650/1849] lr: 4.943e-05, eta: 5 days, 5:48:20, time: 0.492, data_time: 0.024, memory: 4798, loss_cls: 2.0734, loss_bbox: 4.4633, loss_obj: 6.7570, loss: 13.2937 2021-09-01 02:42:41,180 - mmdet - INFO - Epoch [1][700/1849] lr: 5.733e-05, eta: 5 days, 2:08:19, time: 0.485, data_time: 0.026, memory: 4798, loss_cls: 2.1661, loss_bbox: 4.3703, loss_obj: 6.3989, loss: 12.9353 2021-09-01 02:43:05,452 - mmdet - INFO - Epoch [1][750/1849] lr: 6.581e-05, eta: 4 days, 22:57:51, time: 0.485, data_time: 0.024, memory: 4798, loss_cls: 2.1940, loss_bbox: 4.3171, loss_obj: 6.1568, loss: 12.6680 2021-09-01 02:43:30,328 - mmdet - INFO - Epoch [1][800/1849] lr: 7.488e-05, eta: 4 days, 20:18:08, time: 0.497, data_time: 0.024, memory: 4798, loss_cls: 2.2796, loss_bbox: 4.2251, loss_obj: 5.9208, loss: 12.4254 2021-09-01 02:43:54,264 - mmdet - INFO - Epoch [1][850/1849] lr: 8.453e-05, eta: 4 days, 17:46:49, time: 0.478, data_time: 0.026, memory: 4798, loss_cls: 2.3233, loss_bbox: 4.1635, loss_obj: 5.7514, loss: 12.2382 2021-09-01 02:44:18,588 - mmdet - INFO - Epoch [1][900/1849] lr: 9.477e-05, eta: 4 days, 15:36:22, time: 0.486, data_time: 0.026, memory: 4798, loss_cls: 2.3490, loss_bbox: 4.1076, loss_obj: 5.5771, loss: 12.0337 2021-09-01 02:44:42,624 - mmdet - INFO - Epoch [1][950/1849] lr: 1.056e-04, eta: 4 days, 13:37:08, time: 0.481, data_time: 0.026, memory: 4798, loss_cls: 2.3844, loss_bbox: 4.0548, loss_obj: 5.4779, loss: 11.9171 2021-09-01 02:45:06,262 - mmdet - INFO - Exp name: yolox_s_8x8_300e_coco_worker4.py 2021-09-01 02:45:06,263 - mmdet - INFO - Epoch [1][1000/1849] lr: 1.170e-04, eta: 4 days, 11:45:41, time: 0.472, data_time: 0.024, memory: 4798, loss_cls: 2.4346, loss_bbox: 3.9878, loss_obj: 5.3613, loss: 11.7837 2021-09-01 02:45:29,670 - mmdet - INFO - Epoch [1][1050/1849] lr: 1.290e-04, eta: 4 days, 10:02:56, time: 0.468, data_time: 0.025, memory: 4798, loss_cls: 2.4651, loss_bbox: 3.9392, loss_obj: 5.3329, loss: 11.7372 2021-09-01 02:45:54,565 - mmdet - INFO - Epoch [1][1100/1849] lr: 1.416e-04, eta: 4 days, 8:42:10, time: 0.498, data_time: 0.025, memory: 4798, loss_cls: 2.4819, loss_bbox: 3.9040, loss_obj: 5.2847, loss: 11.6706 2021-09-01 02:46:18,028 - mmdet - INFO - Epoch [1][1150/1849] lr: 1.547e-04, eta: 4 days, 7:16:39, time: 0.469, data_time: 0.026, memory: 4798, loss_cls: 2.5091, loss_bbox: 3.8698, loss_obj: 5.2676, loss: 11.6464 2021-09-01 02:46:42,603 - mmdet - INFO - Epoch [1][1200/1849] lr: 1.685e-04, eta: 4 days, 6:06:53, time: 0.492, data_time: 0.023, memory: 4798, loss_cls: 2.5340, loss_bbox: 3.8249, loss_obj: 5.2256, loss: 11.5845 2021-09-01 02:47:06,335 - mmdet - INFO - Epoch [1][1250/1849] lr: 1.828e-04, eta: 4 days, 4:56:21, time: 0.474, data_time: 0.026, memory: 4798, loss_cls: 2.5381, loss_bbox: 3.8085, loss_obj: 5.2429, loss: 11.5896 2021-09-01 02:47:30,399 - mmdet - INFO - Epoch [1][1300/1849] lr: 1.977e-04, eta: 4 days, 3:53:39, time: 0.481, data_time: 0.027, memory: 4798, loss_cls: 2.5707, loss_bbox: 3.7773, loss_obj: 5.2428, loss: 11.5909 2021-09-01 02:47:56,046 - mmdet - INFO - Epoch [1][1350/1849] lr: 2.132e-04, eta: 4 days, 3:06:16, time: 0.513, data_time: 0.029, memory: 4798, loss_cls: 2.5718, loss_bbox: 3.7531, loss_obj: 5.2378, loss: 11.5627 2021-09-01 02:48:20,205 - mmdet - INFO - Epoch [1][1400/1849] lr: 2.293e-04, eta: 4 days, 2:12:36, time: 0.483, data_time: 0.026, memory: 4798, loss_cls: 2.5983, loss_bbox: 3.7283, loss_obj: 5.2257, loss: 11.5523 2021-09-01 02:48:44,729 - mmdet - INFO - Epoch [1][1450/1849] lr: 2.460e-04, eta: 4 days, 1:24:38, time: 0.490, data_time: 0.025, memory: 4798, loss_cls: 2.6041, loss_bbox: 3.7024, loss_obj: 5.2223, loss: 11.5288 2021-09-01 02:49:08,459 - mmdet - INFO - Epoch [1][1500/1849] lr: 2.633e-04, eta: 4 days, 0:35:13, time: 0.475, data_time: 0.025, memory: 4798, loss_cls: 2.6277, loss_bbox: 3.6833, loss_obj: 5.2177, loss: 11.5287 2021-09-01 02:49:33,180 - mmdet - INFO - Epoch [1][1550/1849] lr: 2.811e-04, eta: 3 days, 23:54:54, time: 0.495, data_time: 0.024, memory: 4798, loss_cls: 2.6267, loss_bbox: 3.6606, loss_obj: 5.2363, loss: 11.5236 2021-09-01 02:49:55,826 - mmdet - INFO - Epoch [1][1600/1849] lr: 2.995e-04, eta: 3 days, 23:05:00, time: 0.453, data_time: 0.026, memory: 4798, loss_cls: 2.6458, loss_bbox: 3.6306, loss_obj: 5.2416, loss: 11.5180 2021-09-01 02:50:19,347 - mmdet - INFO - Epoch [1][1650/1849] lr: 3.185e-04, eta: 3 days, 22:23:04, time: 0.471, data_time: 0.025, memory: 4798, loss_cls: 2.6380, loss_bbox: 3.6194, loss_obj: 5.2388, loss: 11.4961 2021-09-01 02:50:43,537 - mmdet - INFO - Epoch [1][1700/1849] lr: 3.381e-04, eta: 3 days, 21:47:03, time: 0.483, data_time: 0.025, memory: 4798, loss_cls: 2.6498, loss_bbox: 3.6019, loss_obj: 5.2407, loss: 11.4924 2021-09-01 02:51:07,227 - mmdet - INFO - Epoch [1][1750/1849] lr: 3.583e-04, eta: 3 days, 21:10:37, time: 0.474, data_time: 0.025, memory: 4798, loss_cls: 2.6598, loss_bbox: 3.5832, loss_obj: 5.2211, loss: 11.4641 2021-09-01 02:51:31,651 - mmdet - INFO - Epoch [1][1800/1849] lr: 3.791e-04, eta: 3 days, 20:39:50, time: 0.488, data_time: 0.025, memory: 4798, loss_cls: 2.6535, loss_bbox: 3.5730, loss_obj: 5.2453, loss: 11.4718

RangiLyu commented 3 years ago

The eta time is not the final training time. And you can see the log above: between the last two lines, the logging time gap is 24 seconds but the eta time gap is 30 min.

RaymondByc commented 3 years ago

But it is a rough estimation for the rest time. Another point is the iter cost time is longer (2x) than the official code

hhaAndroid commented 3 years ago

@RaymondByc Regarding this 2x acceleration implementation, they will load the entire data set into memory. Usually, users don’t have such a large amount of memory.

RaymondByc commented 3 years ago

We do not use the "--cached", and all the setting is same. It still needs a longer time. Did you test your training time with the official code?

hhaAndroid commented 3 years ago

In the same environment, I can post the official and mmdet training time log for you later.

RaymondByc commented 3 years ago

ok. Maybe the difference is based on the capacity of the hardware. Waiting for your log.

RaymondByc commented 3 years ago

Any process?

hhaAndroid commented 3 years ago

Any process?

We will release all YOLOX models and weights recently, you can check the training time through log.

captainIT commented 8 months ago

不仅是yolox 其他模型应该也都可以 https://github.com/NVIDIA/apex/issues/304