从采集数据到部署到JetsonNano全流程
转自AI Studio，原文链接：从采集数据到部署到JetsonNano全流程 - 飞桨AI StudioBML Codelab基于JupyterLab 全新架构升级，支持亮暗主题切换和丰富的AI工具，详见使用说明文档。1. 采集数据集 + labelme标记采集+标注后得到如下所示的 图片+json格式的文件，就是images数据集里的原始数据。2. 整理成coco数据集格式In [20]# 拉
·
转自AI Studio，原文链接：从采集数据到部署到JetsonNano全流程 - 飞桨AI Studio
BML Codelab基于JupyterLab 全新架构升级，支持亮暗主题切换和丰富的AI工具，详见使用说明文档。
1. 采集数据集 + labelme标记
采集+标注后得到如下所示的 图片+json格式的文件，就是images数据集里的原始数据。
2. 整理成coco数据集格式
In [20]
# 拉取PaddleDetection源码,只跑一次
# !git clone https://gitee.com/PaddlePaddle/PaddleDetection
# # 解压数据集，只跑一次
# !unzip -q data/data138474/images.zip -d data/In [23]
# 只运行一次
# import os
# os.chdir('data')
# !python x2coco.py \
# --dataset_type labelme \
# --json_input_dir images\
# --image_input_dir images/ \
# --output_dir ./cocome/ \
# --train_proportion 0.8 \
# --val_proportion 0.2 \
# --test_proportion 0.0
# os.chdir('/home/aistudio')3. PaddleDetection训练
3.1 部署工具只支持以下模型，因此只能选择以下模型中的一个进行训练，我选的是ppyolo
3.2 在PaddleDection/configs/ppyolo 下新建model.yml
_BASE_: [
'../runtime.yml',
'./_base_/ppyolo_tiny.yml',
'./_base_/optimizer_650e.yml',
'./_base_/ppyolo_tiny_reader.yml',
]
snapshot_epoch: 1
weights: output/ppyolo_tiny_650e_coco/model_final
metric: COCO
num_classes: 2
TrainDataset:
!COCODataSet
image_dir: train
anno_path: annotations/instance_train.json
dataset_dir: .././data/cocome
data_fields: ['image', 'gt_bbox', 'gt_class', 'is_crowd']
EvalDataset:
!COCODataSet
image_dir: val
anno_path: annotations/instance_val.json
dataset_dir: .././data/cocome
TestDataset:
!ImageFolder
anno_path: annotations/instance_val.json
dataset_dir: .././data/cocome
#
snapshot_epoch: 5
pretrain_weights: https://paddledet.bj.bcebos.com/models/pretrained/MobileNetV1_ssld_pretrained.pdparams
weights: output/yolov3_mobilenet_v1_ssld_270e_coco/model_final
In [22]
import os
os.chdir('PaddleDetection')
# 只运行一次
# !pip install pycocotools
# !pip install lap
# !pip install motmetrics
!python tools/train.py -c configs/ppyolo/model.yml
os.chdir('/home/aistudio')Warning: import ppdet from source directory without installing, run 'python setup.py install' to install ppdet firstly
loading annotations into memory...
Done (t=0.01s)
creating index...
index created!
W0414 13:29:00.739207 21029 device_context.cc:447] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 10.1, Runtime API Version: 10.1
W0414 13:29:00.744254 21029 device_context.cc:465] device: 0, cuDNN Version: 7.6.
[04/14 13:29:09] ppdet.utils.checkpoint INFO: ['backbone.conv1._batch_norm._mean', 'backbone.conv1._batch_norm._variance', 'backbone.conv1._batch_norm.bias', 'backbone.conv1._batch_norm.weight', 'backbone.conv1._conv.weight', 'backbone.conv2_1._depthwise_conv._batch_norm._mean', 'backbone.conv2_1._depthwise_conv._batch_norm._variance', 'backbone.conv2_1._depthwise_conv._batch_norm.bias', 'backbone.conv2_1._depthwise_conv._batch_norm.weight', 'backbone.conv2_1._depthwise_conv._conv.weight', 'backbone.conv2_1._pointwise_conv._batch_norm._mean', 'backbone.conv2_1._pointwise_conv._batch_norm._variance', 'backbone.conv2_1._pointwise_conv._batch_norm.bias', 'backbone.conv2_1._pointwise_conv._batch_norm.weight', 'backbone.conv2_1._pointwise_conv._conv.weight', 'backbone.conv2_2._depthwise_conv._batch_norm._mean', 'backbone.conv2_2._depthwise_conv._batch_norm._variance', 'backbone.conv2_2._depthwise_conv._batch_norm.bias', 'backbone.conv2_2._depthwise_conv._batch_norm.weight', 'backbone.conv2_2._depthwise_conv._conv.weight', 'backbone.conv2_2._pointwise_conv._batch_norm._mean', 'backbone.conv2_2._pointwise_conv._batch_norm._variance', 'backbone.conv2_2._pointwise_conv._batch_norm.bias', 'backbone.conv2_2._pointwise_conv._batch_norm.weight', 'backbone.conv2_2._pointwise_conv._conv.weight', 'backbone.conv3_1._depthwise_conv._batch_norm._mean', 'backbone.conv3_1._depthwise_conv._batch_norm._variance', 'backbone.conv3_1._depthwise_conv._batch_norm.bias', 'backbone.conv3_1._depthwise_conv._batch_norm.weight', 'backbone.conv3_1._depthwise_conv._conv.weight', 'backbone.conv3_1._pointwise_conv._batch_norm._mean', 'backbone.conv3_1._pointwise_conv._batch_norm._variance', 'backbone.conv3_1._pointwise_conv._batch_norm.bias', 'backbone.conv3_1._pointwise_conv._batch_norm.weight', 'backbone.conv3_1._pointwise_conv._conv.weight', 'backbone.conv3_2._depthwise_conv._batch_norm._mean', 'backbone.conv3_2._depthwise_conv._batch_norm._variance', 'backbone.conv3_2._depthwise_conv._batch_norm.bias', 'backbone.conv3_2._depthwise_conv._batch_norm.weight', 'backbone.conv3_2._depthwise_conv._conv.weight', 'backbone.conv3_2._pointwise_conv._batch_norm._mean', 'backbone.conv3_2._pointwise_conv._batch_norm._variance', 'backbone.conv3_2._pointwise_conv._batch_norm.bias', 'backbone.conv3_2._pointwise_conv._batch_norm.weight', 'backbone.conv3_2._pointwise_conv._conv.weight', 'backbone.conv4_1._depthwise_conv._batch_norm._mean', 'backbone.conv4_1._depthwise_conv._batch_norm._variance', 'backbone.conv4_1._depthwise_conv._batch_norm.bias', 'backbone.conv4_1._depthwise_conv._batch_norm.weight', 'backbone.conv4_1._depthwise_conv._conv.weight', 'backbone.conv4_1._pointwise_conv._batch_norm._mean', 'backbone.conv4_1._pointwise_conv._batch_norm._variance', 'backbone.conv4_1._pointwise_conv._batch_norm.bias', 'backbone.conv4_1._pointwise_conv._batch_norm.weight', 'backbone.conv4_1._pointwise_conv._conv.weight', 'backbone.conv4_2._depthwise_conv._batch_norm._mean', 'backbone.conv4_2._depthwise_conv._batch_norm._variance', 'backbone.conv4_2._depthwise_conv._batch_norm.bias', 'backbone.conv4_2._depthwise_conv._batch_norm.weight', 'backbone.conv4_2._depthwise_conv._conv.weight', 'backbone.conv4_2._pointwise_conv._batch_norm._mean', 'backbone.conv4_2._pointwise_conv._batch_norm._variance', 'backbone.conv4_2._pointwise_conv._batch_norm.bias', 'backbone.conv4_2._pointwise_conv._batch_norm.weight', 'backbone.conv4_2._pointwise_conv._conv.weight', 'backbone.conv5_1._depthwise_conv._batch_norm._mean', 'backbone.conv5_1._depthwise_conv._batch_norm._variance', 'backbone.conv5_1._depthwise_conv._batch_norm.bias', 'backbone.conv5_1._depthwise_conv._batch_norm.weight', 'backbone.conv5_1._depthwise_conv._conv.weight', 'backbone.conv5_1._pointwise_conv._batch_norm._mean', 'backbone.conv5_1._pointwise_conv._batch_norm._variance', 'backbone.conv5_1._pointwise_conv._batch_norm.bias', 'backbone.conv5_1._pointwise_conv._batch_norm.weight', 'backbone.conv5_1._pointwise_conv._conv.weight', 'backbone.conv5_2._depthwise_conv._batch_norm._mean', 'backbone.conv5_2._depthwise_conv._batch_norm._variance', 'backbone.conv5_2._depthwise_conv._batch_norm.bias', 'backbone.conv5_2._depthwise_conv._batch_norm.weight', 'backbone.conv5_2._depthwise_conv._conv.weight', 'backbone.conv5_2._pointwise_conv._batch_norm._mean', 'backbone.conv5_2._pointwise_conv._batch_norm._variance', 'backbone.conv5_2._pointwise_conv._batch_norm.bias', 'backbone.conv5_2._pointwise_conv._batch_norm.weight', 'backbone.conv5_2._pointwise_conv._conv.weight', 'backbone.conv5_3._depthwise_conv._batch_norm._mean', 'backbone.conv5_3._depthwise_conv._batch_norm._variance', 'backbone.conv5_3._depthwise_conv._batch_norm.bias', 'backbone.conv5_3._depthwise_conv._batch_norm.weight', 'backbone.conv5_3._depthwise_conv._conv.weight', 'backbone.conv5_3._pointwise_conv._batch_norm._mean', 'backbone.conv5_3._pointwise_conv._batch_norm._variance', 'backbone.conv5_3._pointwise_conv._batch_norm.bias', 'backbone.conv5_3._pointwise_conv._batch_norm.weight', 'backbone.conv5_3._pointwise_conv._conv.weight', 'backbone.conv5_4._depthwise_conv._batch_norm._mean', 'backbone.conv5_4._depthwise_conv._batch_norm._variance', 'backbone.conv5_4._depthwise_conv._batch_norm.bias', 'backbone.conv5_4._depthwise_conv._batch_norm.weight', 'backbone.conv5_4._depthwise_conv._conv.weight', 'backbone.conv5_4._pointwise_conv._batch_norm._mean', 'backbone.conv5_4._pointwise_conv._batch_norm._variance', 'backbone.conv5_4._pointwise_conv._batch_norm.bias', 'backbone.conv5_4._pointwise_conv._batch_norm.weight', 'backbone.conv5_4._pointwise_conv._conv.weight', 'backbone.conv5_5._depthwise_conv._batch_norm._mean', 'backbone.conv5_5._depthwise_conv._batch_norm._variance', 'backbone.conv5_5._depthwise_conv._batch_norm.bias', 'backbone.conv5_5._depthwise_conv._batch_norm.weight', 'backbone.conv5_5._depthwise_conv._conv.weight', 'backbone.conv5_5._pointwise_conv._batch_norm._mean', 'backbone.conv5_5._pointwise_conv._batch_norm._variance', 'backbone.conv5_5._pointwise_conv._batch_norm.bias', 'backbone.conv5_5._pointwise_conv._batch_norm.weight', 'backbone.conv5_5._pointwise_conv._conv.weight', 'backbone.conv5_6._depthwise_conv._batch_norm._mean', 'backbone.conv5_6._depthwise_conv._batch_norm._variance', 'backbone.conv5_6._depthwise_conv._batch_norm.bias', 'backbone.conv5_6._depthwise_conv._batch_norm.weight', 'backbone.conv5_6._depthwise_conv._conv.weight', 'backbone.conv5_6._pointwise_conv._batch_norm._mean', 'backbone.conv5_6._pointwise_conv._batch_norm._variance', 'backbone.conv5_6._pointwise_conv._batch_norm.bias', 'backbone.conv5_6._pointwise_conv._batch_norm.weight', 'backbone.conv5_6._pointwise_conv._conv.weight', 'backbone.conv6._depthwise_conv._batch_norm._mean', 'backbone.conv6._depthwise_conv._batch_norm._variance', 'backbone.conv6._depthwise_conv._batch_norm.bias', 'backbone.conv6._depthwise_conv._batch_norm.weight', 'backbone.conv6._depthwise_conv._conv.weight', 'backbone.conv6._pointwise_conv._batch_norm._mean', 'backbone.conv6._pointwise_conv._batch_norm._variance', 'backbone.conv6._pointwise_conv._batch_norm.bias', 'backbone.conv6._pointwise_conv._batch_norm.weight', 'backbone.conv6._pointwise_conv._conv.weight'] in pretrained weight is not used in the model, and its will not be loaded
[04/14 13:29:09] ppdet.utils.checkpoint INFO: Finish loading model weights: /home/aistudio/.cache/paddle/weights/MobileNetV1_ssld_pretrained.pdparams
[04/14 13:29:09] ppdet.engine INFO: Epoch: [0] [ 0/18] learning_rate: 0.000000 loss_xy: 1.586861 loss_wh: 6.782475 loss_iou: 5.943724 loss_obj: 2636.933594 loss_cls: 2.026736 loss: 2653.273438 eta: 0:45:13 batch_cost: 0.2319 data_cost: 0.0004 ips: 138.0005 images/s
[04/14 13:29:25] ppdet.engine INFO: Epoch: [1] [ 0/18] learning_rate: 0.000022 loss_xy: 1.670596 loss_wh: 6.782475 loss_iou: 6.184143 loss_obj: 1738.167725 loss_cls: 2.277970 loss: 1752.807739 eta: 2:21:48 batch_cost: 0.7283 data_cost: 0.2621 ips: 43.9371 images/s
[04/14 13:29:45] ppdet.engine INFO: Epoch: [2] [ 0/18] learning_rate: 0.000045 loss_xy: 1.652527 loss_wh: 5.740756 loss_iou: 6.207388 loss_obj: 28.590870 loss_cls: 2.235853 loss: 43.590027 eta: 2:38:10 batch_cost: 1.0318 data_cost: 0.5905 ips: 31.0123 images/s
[04/14 13:30:03] ppdet.engine INFO: Epoch: [3] [ 0/18] learning_rate: 0.000068 loss_xy: 1.677333 loss_wh: 4.489705 loss_iou: 6.079871 loss_obj: 16.190971 loss_cls: 2.188293 loss: 30.777775 eta: 2:41:15 batch_cost: 0.9597 data_cost: 0.5339 ips: 33.3454 images/s
[04/14 13:30:22] ppdet.engine INFO: Epoch: [4] [ 0/18] learning_rate: 0.000090 loss_xy: 1.586078 loss_wh: 3.681612 loss_iou: 5.730135 loss_obj: 14.666073 loss_cls: 2.115413 loss: 27.438993 eta: 2:44:31 batch_cost: 1.0519 data_cost: 0.6384 ips: 30.4215 images/s
[04/14 13:30:37] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:30:41] ppdet.engine INFO: Epoch: [5] [ 0/18] learning_rate: 0.000112 loss_xy: 1.506183 loss_wh: 3.137302 loss_iou: 5.552191 loss_obj: 13.259188 loss_cls: 2.053119 loss: 25.142277 eta: 2:45:04 batch_cost: 1.0160 data_cost: 0.6120 ips: 31.4972 images/s
[04/14 13:31:01] ppdet.engine INFO: Epoch: [6] [ 0/18] learning_rate: 0.000135 loss_xy: 1.455660 loss_wh: 2.865020 loss_iou: 5.466364 loss_obj: 12.259902 loss_cls: 2.015674 loss: 24.021690 eta: 2:47:59 batch_cost: 1.0866 data_cost: 0.6491 ips: 29.4492 images/s
[04/14 13:31:20] ppdet.engine INFO: Epoch: [7] [ 0/18] learning_rate: 0.000158 loss_xy: 1.442592 loss_wh: 2.355444 loss_iou: 5.155467 loss_obj: 11.294309 loss_cls: 1.946530 loss: 22.355350 eta: 2:47:41 batch_cost: 0.9832 data_cost: 0.5852 ips: 32.5476 images/s
[04/14 13:31:37] ppdet.engine INFO: Epoch: [8] [ 0/18] learning_rate: 0.000180 loss_xy: 1.519472 loss_wh: 2.367092 loss_iou: 5.155847 loss_obj: 11.428535 loss_cls: 2.016512 loss: 22.397797 eta: 2:46:24 batch_cost: 0.9573 data_cost: 0.5725 ips: 33.4289 images/s
[04/14 13:31:56] ppdet.engine INFO: Epoch: [9] [ 0/18] learning_rate: 0.000203 loss_xy: 1.497352 loss_wh: 2.189704 loss_iou: 5.137483 loss_obj: 10.422691 loss_cls: 1.907944 loss: 21.197552 eta: 2:46:43 batch_cost: 1.0113 data_cost: 0.6285 ips: 31.6421 images/s
[04/14 13:32:13] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:32:16] ppdet.engine INFO: Epoch: [10] [ 0/18] learning_rate: 0.000225 loss_xy: 1.487687 loss_wh: 2.089318 loss_iou: 5.031182 loss_obj: 10.473335 loss_cls: 1.901937 loss: 21.044270 eta: 2:46:39 batch_cost: 1.0446 data_cost: 0.5962 ips: 30.6329 images/s
[04/14 13:32:35] ppdet.engine INFO: Epoch: [11] [ 0/18] learning_rate: 0.000247 loss_xy: 1.445958 loss_wh: 2.005820 loss_iou: 5.001819 loss_obj: 10.060132 loss_cls: 1.875436 loss: 20.637432 eta: 2:47:09 batch_cost: 0.9941 data_cost: 0.5988 ips: 32.1903 images/s
[04/14 13:32:54] ppdet.engine INFO: Epoch: [12] [ 0/18] learning_rate: 0.000270 loss_xy: 1.506358 loss_wh: 1.877072 loss_iou: 4.918931 loss_obj: 9.635354 loss_cls: 1.871274 loss: 20.051880 eta: 2:47:46 batch_cost: 1.0575 data_cost: 0.6675 ips: 30.2590 images/s
[04/14 13:33:13] ppdet.engine INFO: Epoch: [13] [ 0/18] learning_rate: 0.000293 loss_xy: 1.465201 loss_wh: 1.725131 loss_iou: 4.792963 loss_obj: 9.460852 loss_cls: 1.865794 loss: 19.288349 eta: 2:48:07 batch_cost: 1.0288 data_cost: 0.6296 ips: 31.1035 images/s
[04/14 13:33:31] ppdet.engine INFO: Epoch: [14] [ 0/18] learning_rate: 0.000315 loss_xy: 1.428189 loss_wh: 1.685681 loss_iou: 4.628690 loss_obj: 9.353590 loss_cls: 1.822749 loss: 18.887215 eta: 2:47:25 batch_cost: 1.0511 data_cost: 0.6709 ips: 30.4456 images/s
[04/14 13:33:47] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:33:50] ppdet.engine INFO: Epoch: [15] [ 0/18] learning_rate: 0.000338 loss_xy: 1.361993 loss_wh: 1.557696 loss_iou: 4.563309 loss_obj: 9.053027 loss_cls: 1.824860 loss: 18.748085 eta: 2:46:09 batch_cost: 0.9061 data_cost: 0.4945 ips: 35.3166 images/s
[04/14 13:34:09] ppdet.engine INFO: Epoch: [16] [ 0/18] learning_rate: 0.000360 loss_xy: 1.379558 loss_wh: 1.499877 loss_iou: 4.459455 loss_obj: 8.726969 loss_cls: 1.763658 loss: 18.001774 eta: 2:46:05 batch_cost: 0.9612 data_cost: 0.5154 ips: 33.2908 images/s
[04/14 13:34:29] ppdet.engine INFO: Epoch: [17] [ 0/18] learning_rate: 0.000383 loss_xy: 1.378514 loss_wh: 1.547090 loss_iou: 4.644367 loss_obj: 8.953167 loss_cls: 1.801022 loss: 18.301060 eta: 2:46:39 batch_cost: 1.0928 data_cost: 0.6452 ips: 29.2813 images/s
[04/14 13:34:49] ppdet.engine INFO: Epoch: [18] [ 0/18] learning_rate: 0.000405 loss_xy: 1.364754 loss_wh: 1.419811 loss_iou: 4.493394 loss_obj: 8.574642 loss_cls: 1.757815 loss: 17.496799 eta: 2:46:48 batch_cost: 0.9995 data_cost: 0.6053 ips: 32.0167 images/s
[04/14 13:35:07] ppdet.engine INFO: Epoch: [19] [ 0/18] learning_rate: 0.000428 loss_xy: 1.341500 loss_wh: 1.377572 loss_iou: 4.375639 loss_obj: 8.022249 loss_cls: 1.748245 loss: 16.990774 eta: 2:46:21 batch_cost: 1.0478 data_cost: 0.6480 ips: 30.5399 images/s
[04/14 13:35:21] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:35:25] ppdet.engine INFO: Epoch: [20] [ 0/18] learning_rate: 0.000450 loss_xy: 1.348360 loss_wh: 1.395988 loss_iou: 4.533930 loss_obj: 7.927083 loss_cls: 1.741763 loss: 17.198647 eta: 2:45:45 batch_cost: 0.9554 data_cost: 0.6301 ips: 33.4936 images/s
[04/14 13:35:46] ppdet.engine INFO: Epoch: [21] [ 0/18] learning_rate: 0.000472 loss_xy: 1.347176 loss_wh: 1.315490 loss_iou: 4.399605 loss_obj: 7.825877 loss_cls: 1.731823 loss: 16.570663 eta: 2:46:29 batch_cost: 1.0895 data_cost: 0.6902 ips: 29.3725 images/s
[04/14 13:36:06] ppdet.engine INFO: Epoch: [22] [ 0/18] learning_rate: 0.000495 loss_xy: 1.279625 loss_wh: 1.321942 loss_iou: 4.124500 loss_obj: 7.821168 loss_cls: 1.665033 loss: 16.358036 eta: 2:46:39 batch_cost: 1.0822 data_cost: 0.6581 ips: 29.5689 images/s
[04/14 13:36:22] ppdet.engine INFO: Epoch: [23] [ 0/18] learning_rate: 0.000517 loss_xy: 1.283677 loss_wh: 1.324662 loss_iou: 4.344380 loss_obj: 8.136177 loss_cls: 1.699326 loss: 16.736591 eta: 2:45:28 batch_cost: 1.0078 data_cost: 0.6475 ips: 31.7532 images/s
[04/14 13:36:44] ppdet.engine INFO: Epoch: [24] [ 0/18] learning_rate: 0.000540 loss_xy: 1.279502 loss_wh: 1.260314 loss_iou: 4.317026 loss_obj: 7.755419 loss_cls: 1.718941 loss: 16.246592 eta: 2:46:13 batch_cost: 1.0751 data_cost: 0.6860 ips: 29.7639 images/s
[04/14 13:36:59] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:37:01] ppdet.engine INFO: Epoch: [25] [ 0/18] learning_rate: 0.000563 loss_xy: 1.339955 loss_wh: 1.263783 loss_iou: 4.324253 loss_obj: 7.587839 loss_cls: 1.697272 loss: 16.153664 eta: 2:45:04 batch_cost: 0.9771 data_cost: 0.5395 ips: 32.7505 images/s
[04/14 13:37:21] ppdet.engine INFO: Epoch: [26] [ 0/18] learning_rate: 0.000585 loss_xy: 1.323967 loss_wh: 1.237977 loss_iou: 4.300310 loss_obj: 7.414727 loss_cls: 1.713512 loss: 15.955509 eta: 2:45:18 batch_cost: 0.9859 data_cost: 0.5975 ips: 32.4566 images/s
[04/14 13:37:39] ppdet.engine INFO: Epoch: [27] [ 0/18] learning_rate: 0.000607 loss_xy: 1.299361 loss_wh: 1.162107 loss_iou: 4.169978 loss_obj: 7.097247 loss_cls: 1.653652 loss: 15.428999 eta: 2:44:55 batch_cost: 1.0013 data_cost: 0.6169 ips: 31.9583 images/s
[04/14 13:37:58] ppdet.engine INFO: Epoch: [28] [ 0/18] learning_rate: 0.000630 loss_xy: 1.219962 loss_wh: 1.163996 loss_iou: 4.114856 loss_obj: 6.879937 loss_cls: 1.623702 loss: 14.862214 eta: 2:44:42 batch_cost: 1.0693 data_cost: 0.6371 ips: 29.9257 images/s
[04/14 13:38:17] ppdet.engine INFO: Epoch: [29] [ 0/18] learning_rate: 0.000653 loss_xy: 1.288635 loss_wh: 1.188367 loss_iou: 4.203623 loss_obj: 6.869554 loss_cls: 1.658932 loss: 15.210352 eta: 2:44:28 batch_cost: 1.0212 data_cost: 0.6037 ips: 31.3350 images/s
[04/14 13:38:30] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:38:34] ppdet.engine INFO: Epoch: [30] [ 0/18] learning_rate: 0.000675 loss_xy: 1.242197 loss_wh: 1.162710 loss_iou: 4.124510 loss_obj: 6.521726 loss_cls: 1.625620 loss: 14.792933 eta: 2:43:27 batch_cost: 0.9237 data_cost: 0.5475 ips: 34.6441 images/s
[04/14 13:38:52] ppdet.engine INFO: Epoch: [31] [ 0/18] learning_rate: 0.000698 loss_xy: 1.302225 loss_wh: 1.148133 loss_iou: 4.210639 loss_obj: 6.531621 loss_cls: 1.716509 loss: 15.117219 eta: 2:42:43 batch_cost: 0.9170 data_cost: 0.5760 ips: 34.8982 images/s
[04/14 13:39:11] ppdet.engine INFO: Epoch: [32] [ 0/18] learning_rate: 0.000720 loss_xy: 1.213626 loss_wh: 1.134214 loss_iou: 4.123302 loss_obj: 6.632785 loss_cls: 1.599706 loss: 14.834019 eta: 2:42:42 batch_cost: 1.0483 data_cost: 0.6387 ips: 30.5250 images/s
[04/14 13:39:28] ppdet.engine INFO: Epoch: [33] [ 0/18] learning_rate: 0.000742 loss_xy: 1.175269 loss_wh: 1.106271 loss_iou: 3.905134 loss_obj: 6.463820 loss_cls: 1.650248 loss: 14.430290 eta: 2:42:01 batch_cost: 0.9255 data_cost: 0.5401 ips: 34.5773 images/s
[04/14 13:39:47] ppdet.engine INFO: Epoch: [34] [ 0/18] learning_rate: 0.000765 loss_xy: 1.192600 loss_wh: 1.136012 loss_iou: 4.224728 loss_obj: 6.844395 loss_cls: 1.669614 loss: 14.857993 eta: 2:41:44 batch_cost: 0.9511 data_cost: 0.5711 ips: 33.6459 images/s
[04/14 13:40:03] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:40:07] ppdet.engine INFO: Epoch: [35] [ 0/18] learning_rate: 0.000788 loss_xy: 1.155041 loss_wh: 1.098137 loss_iou: 4.002364 loss_obj: 6.452302 loss_cls: 1.632790 loss: 14.179432 eta: 2:41:46 batch_cost: 1.0367 data_cost: 0.6305 ips: 30.8683 images/s
[04/14 13:40:28] ppdet.engine INFO: Epoch: [36] [ 0/18] learning_rate: 0.000810 loss_xy: 1.206504 loss_wh: 1.132179 loss_iou: 4.116858 loss_obj: 6.299964 loss_cls: 1.648950 loss: 14.581146 eta: 2:41:51 batch_cost: 1.0439 data_cost: 0.6548 ips: 30.6535 images/s
[04/14 13:40:47] ppdet.engine INFO: Epoch: [37] [ 0/18] learning_rate: 0.000833 loss_xy: 1.210473 loss_wh: 1.071723 loss_iou: 4.017380 loss_obj: 6.162838 loss_cls: 1.668937 loss: 14.363120 eta: 2:41:41 batch_cost: 1.0995 data_cost: 0.7148 ips: 29.1046 images/s
[04/14 13:41:06] ppdet.engine INFO: Epoch: [38] [ 0/18] learning_rate: 0.000855 loss_xy: 1.162920 loss_wh: 1.135330 loss_iou: 4.008078 loss_obj: 5.982968 loss_cls: 1.642006 loss: 14.023290 eta: 2:41:40 batch_cost: 1.0553 data_cost: 0.6463 ips: 30.3232 images/s
[04/14 13:41:24] ppdet.engine INFO: Epoch: [39] [ 0/18] learning_rate: 0.000877 loss_xy: 1.180532 loss_wh: 1.068939 loss_iou: 4.020936 loss_obj: 6.100730 loss_cls: 1.589628 loss: 14.088037 eta: 2:41:11 batch_cost: 0.9979 data_cost: 0.6116 ips: 32.0678 images/s
[04/14 13:41:40] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:41:44] ppdet.engine INFO: Epoch: [40] [ 0/18] learning_rate: 0.000900 loss_xy: 1.182320 loss_wh: 1.001981 loss_iou: 3.832108 loss_obj: 5.905117 loss_cls: 1.591098 loss: 13.336836 eta: 2:41:05 batch_cost: 1.0317 data_cost: 0.6385 ips: 31.0174 images/s
[04/14 13:42:02] ppdet.engine INFO: Epoch: [41] [ 0/18] learning_rate: 0.000923 loss_xy: 1.105936 loss_wh: 1.059723 loss_iou: 3.831776 loss_obj: 5.905568 loss_cls: 1.595631 loss: 13.585750 eta: 2:40:38 batch_cost: 0.9647 data_cost: 0.5491 ips: 33.1724 images/s
[04/14 13:42:20] ppdet.engine INFO: Epoch: [42] [ 0/18] learning_rate: 0.000945 loss_xy: 1.185375 loss_wh: 1.039340 loss_iou: 3.797884 loss_obj: 5.847095 loss_cls: 1.497744 loss: 13.561184 eta: 2:40:17 batch_cost: 1.0292 data_cost: 0.6239 ips: 31.0929 images/s
[04/14 13:42:40] ppdet.engine INFO: Epoch: [43] [ 0/18] learning_rate: 0.000968 loss_xy: 1.173368 loss_wh: 1.017938 loss_iou: 3.803827 loss_obj: 5.726425 loss_cls: 1.546683 loss: 13.308468 eta: 2:40:15 batch_cost: 1.0556 data_cost: 0.6405 ips: 30.3146 images/s
[04/14 13:42:59] ppdet.engine INFO: Epoch: [44] [ 0/18] learning_rate: 0.000990 loss_xy: 1.131293 loss_wh: 1.026389 loss_iou: 3.750814 loss_obj: 5.461053 loss_cls: 1.497626 loss: 13.194307 eta: 2:40:06 batch_cost: 1.0082 data_cost: 0.6113 ips: 31.7392 images/s
[04/14 13:43:14] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:43:17] ppdet.engine INFO: Epoch: [45] [ 0/18] learning_rate: 0.001013 loss_xy: 1.127260 loss_wh: 1.043500 loss_iou: 3.881945 loss_obj: 5.401156 loss_cls: 1.527612 loss: 13.098871 eta: 2:39:30 batch_cost: 0.9761 data_cost: 0.5771 ips: 32.7842 images/s
[04/14 13:43:37] ppdet.engine INFO: Epoch: [46] [ 0/18] learning_rate: 0.001035 loss_xy: 1.098794 loss_wh: 0.993629 loss_iou: 3.699533 loss_obj: 5.536014 loss_cls: 1.487962 loss: 12.846670 eta: 2:39:35 batch_cost: 1.0171 data_cost: 0.6093 ips: 31.4619 images/s
[04/14 13:43:55] ppdet.engine INFO: Epoch: [47] [ 0/18] learning_rate: 0.001058 loss_xy: 1.088106 loss_wh: 0.971704 loss_iou: 3.746615 loss_obj: 5.330320 loss_cls: 1.475442 loss: 12.465189 eta: 2:39:08 batch_cost: 0.9884 data_cost: 0.6141 ips: 32.3755 images/s
[04/14 13:44:13] ppdet.engine INFO: Epoch: [48] [ 0/18] learning_rate: 0.001080 loss_xy: 1.126358 loss_wh: 1.041296 loss_iou: 3.877872 loss_obj: 5.504457 loss_cls: 1.521790 loss: 13.025694 eta: 2:38:48 batch_cost: 0.9964 data_cost: 0.6379 ips: 32.1164 images/s
[04/14 13:44:32] ppdet.engine INFO: Epoch: [49] [ 0/18] learning_rate: 0.001102 loss_xy: 1.137924 loss_wh: 0.997383 loss_iou: 3.795869 loss_obj: 5.227892 loss_cls: 1.460483 loss: 12.519150 eta: 2:38:45 batch_cost: 1.0309 data_cost: 0.6487 ips: 31.0411 images/s
[04/14 13:44:48] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:44:52] ppdet.engine INFO: Epoch: [50] [ 0/18] learning_rate: 0.001125 loss_xy: 1.123474 loss_wh: 1.007330 loss_iou: 3.760498 loss_obj: 5.437192 loss_cls: 1.382188 loss: 12.624385 eta: 2:38:26 batch_cost: 1.0575 data_cost: 0.6440 ips: 30.2611 images/s
[04/14 13:45:09] ppdet.engine INFO: Epoch: [51] [ 0/18] learning_rate: 0.001148 loss_xy: 1.088140 loss_wh: 1.034683 loss_iou: 3.883729 loss_obj: 5.246078 loss_cls: 1.417699 loss: 12.534090 eta: 2:38:01 batch_cost: 0.9790 data_cost: 0.5878 ips: 32.6871 images/s
[04/14 13:45:28] ppdet.engine INFO: Epoch: [52] [ 0/18] learning_rate: 0.001170 loss_xy: 1.087849 loss_wh: 1.007386 loss_iou: 3.641879 loss_obj: 4.800151 loss_cls: 1.375220 loss: 12.029517 eta: 2:37:38 batch_cost: 0.9236 data_cost: 0.5511 ips: 34.6452 images/s
[04/14 13:45:46] ppdet.engine INFO: Epoch: [53] [ 0/18] learning_rate: 0.001192 loss_xy: 1.083071 loss_wh: 1.026600 loss_iou: 3.700399 loss_obj: 5.068700 loss_cls: 1.373246 loss: 12.185109 eta: 2:37:21 batch_cost: 0.9741 data_cost: 0.5783 ips: 32.8508 images/s
[04/14 13:46:07] ppdet.engine INFO: Epoch: [54] [ 0/18] learning_rate: 0.001215 loss_xy: 1.063921 loss_wh: 0.995095 loss_iou: 3.728944 loss_obj: 5.028282 loss_cls: 1.402343 loss: 12.247927 eta: 2:37:29 batch_cost: 1.1036 data_cost: 0.6622 ips: 28.9966 images/s
[04/14 13:46:22] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:46:25] ppdet.engine INFO: Epoch: [55] [ 0/18] learning_rate: 0.001238 loss_xy: 1.093541 loss_wh: 0.993874 loss_iou: 3.784433 loss_obj: 5.018102 loss_cls: 1.352228 loss: 12.268488 eta: 2:36:53 batch_cost: 0.9333 data_cost: 0.5571 ips: 34.2860 images/s
[04/14 13:46:43] ppdet.engine INFO: Epoch: [56] [ 0/18] learning_rate: 0.001260 loss_xy: 1.032600 loss_wh: 1.028335 loss_iou: 3.830704 loss_obj: 5.086860 loss_cls: 1.346473 loss: 12.154493 eta: 2:36:33 batch_cost: 0.9429 data_cost: 0.5473 ips: 33.9377 images/s
[04/14 13:47:04] ppdet.engine INFO: Epoch: [57] [ 0/18] learning_rate: 0.001283 loss_xy: 1.063488 loss_wh: 0.972981 loss_iou: 3.562591 loss_obj: 4.622534 loss_cls: 1.274776 loss: 11.995181 eta: 2:36:42 batch_cost: 1.0857 data_cost: 0.6603 ips: 29.4746 images/s
[04/14 13:47:23] ppdet.engine INFO: Epoch: [58] [ 0/18] learning_rate: 0.001305 loss_xy: 1.036786 loss_wh: 0.954412 loss_iou: 3.677787 loss_obj: 4.706769 loss_cls: 1.306023 loss: 11.760216 eta: 2:36:18 batch_cost: 0.9890 data_cost: 0.5783 ips: 32.3556 images/s
[04/14 13:47:41] ppdet.engine INFO: Epoch: [59] [ 0/18] learning_rate: 0.001328 loss_xy: 1.040621 loss_wh: 0.988029 loss_iou: 3.617022 loss_obj: 4.638086 loss_cls: 1.280257 loss: 11.549982 eta: 2:36:05 batch_cost: 1.0135 data_cost: 0.6260 ips: 31.5750 images/s
[04/14 13:47:56] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:47:59] ppdet.engine INFO: Epoch: [60] [ 0/18] learning_rate: 0.001350 loss_xy: 1.052227 loss_wh: 1.057238 loss_iou: 3.730057 loss_obj: 4.873676 loss_cls: 1.306327 loss: 12.172830 eta: 2:35:30 batch_cost: 0.9169 data_cost: 0.5374 ips: 34.9013 images/s
[04/14 13:48:19] ppdet.engine INFO: Epoch: [61] [ 0/18] learning_rate: 0.001373 loss_xy: 1.050290 loss_wh: 1.019307 loss_iou: 3.664206 loss_obj: 4.816324 loss_cls: 1.324188 loss: 11.893969 eta: 2:35:24 batch_cost: 1.0137 data_cost: 0.5907 ips: 31.5662 images/s
[04/14 13:48:38] ppdet.engine INFO: Epoch: [62] [ 0/18] learning_rate: 0.001395 loss_xy: 1.014099 loss_wh: 0.929005 loss_iou: 3.509603 loss_obj: 4.703539 loss_cls: 1.202885 loss: 11.171493 eta: 2:35:17 batch_cost: 1.0907 data_cost: 0.6796 ips: 29.3378 images/s
[04/14 13:48:55] ppdet.engine INFO: Epoch: [63] [ 0/18] learning_rate: 0.001417 loss_xy: 1.034544 loss_wh: 0.944866 loss_iou: 3.658864 loss_obj: 4.432128 loss_cls: 1.269261 loss: 11.498484 eta: 2:34:42 batch_cost: 0.9497 data_cost: 0.6232 ips: 33.6956 images/s
[04/14 13:49:16] ppdet.engine INFO: Epoch: [64] [ 0/18] learning_rate: 0.001440 loss_xy: 0.994492 loss_wh: 0.903282 loss_iou: 3.615134 loss_obj: 4.524835 loss_cls: 1.179117 loss: 11.206528 eta: 2:34:49 batch_cost: 1.1257 data_cost: 0.7050 ips: 28.4258 images/s
[04/14 13:49:31] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:49:34] ppdet.engine INFO: Epoch: [65] [ 0/18] learning_rate: 0.001463 loss_xy: 1.011091 loss_wh: 0.941515 loss_iou: 3.596971 loss_obj: 4.656765 loss_cls: 1.148330 loss: 11.386809 eta: 2:34:27 batch_cost: 1.0783 data_cost: 0.6678 ips: 29.6756 images/s
[04/14 13:49:52] ppdet.engine INFO: Epoch: [66] [ 0/18] learning_rate: 0.001485 loss_xy: 0.996215 loss_wh: 0.956152 loss_iou: 3.540395 loss_obj: 4.397746 loss_cls: 1.210909 loss: 11.109131 eta: 2:34:01 batch_cost: 0.8828 data_cost: 0.4966 ips: 36.2472 images/s
[04/14 13:50:10] ppdet.engine INFO: Epoch: [67] [ 0/18] learning_rate: 0.001507 loss_xy: 1.025830 loss_wh: 0.951223 loss_iou: 3.621074 loss_obj: 4.450707 loss_cls: 1.162391 loss: 11.410133 eta: 2:33:45 batch_cost: 0.9857 data_cost: 0.6016 ips: 32.4638 images/s
[04/14 13:50:29] ppdet.engine INFO: Epoch: [68] [ 0/18] learning_rate: 0.001530 loss_xy: 0.975528 loss_wh: 0.920183 loss_iou: 3.397571 loss_obj: 4.251836 loss_cls: 1.078805 loss: 10.679556 eta: 2:33:32 batch_cost: 1.0300 data_cost: 0.6089 ips: 31.0673 images/s
[04/14 13:50:49] ppdet.engine INFO: Epoch: [69] [ 0/18] learning_rate: 0.001553 loss_xy: 0.983081 loss_wh: 0.989966 loss_iou: 3.710008 loss_obj: 4.634967 loss_cls: 1.154060 loss: 11.457384 eta: 2:33:21 batch_cost: 1.0791 data_cost: 0.6648 ips: 29.6537 images/s
[04/14 13:51:04] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:51:08] ppdet.engine INFO: Epoch: [70] [ 0/18] learning_rate: 0.001575 loss_xy: 0.986857 loss_wh: 0.964469 loss_iou: 3.561439 loss_obj: 4.492888 loss_cls: 1.132704 loss: 10.935749 eta: 2:33:03 batch_cost: 0.9978 data_cost: 0.6053 ips: 32.0716 images/s
[04/14 13:51:25] ppdet.engine INFO: Epoch: [71] [ 0/18] learning_rate: 0.001597 loss_xy: 0.991727 loss_wh: 0.919071 loss_iou: 3.489734 loss_obj: 4.451149 loss_cls: 1.159353 loss: 10.867794 eta: 2:32:34 batch_cost: 0.9585 data_cost: 0.5851 ips: 33.3869 images/s
[04/14 13:51:44] ppdet.engine INFO: Epoch: [72] [ 0/18] learning_rate: 0.001620 loss_xy: 0.992301 loss_wh: 0.910219 loss_iou: 3.636572 loss_obj: 4.278091 loss_cls: 1.138956 loss: 10.937656 eta: 2:32:24 batch_cost: 1.0216 data_cost: 0.6478 ips: 31.3234 images/s
[04/14 13:52:03] ppdet.engine INFO: Epoch: [73] [ 0/18] learning_rate: 0.001643 loss_xy: 1.002902 loss_wh: 0.960504 loss_iou: 3.694657 loss_obj: 4.276593 loss_cls: 1.136110 loss: 11.131372 eta: 2:32:05 batch_cost: 1.0153 data_cost: 0.6098 ips: 31.5187 images/s
[04/14 13:52:21] ppdet.engine INFO: Epoch: [74] [ 0/18] learning_rate: 0.001665 loss_xy: 0.941965 loss_wh: 0.894139 loss_iou: 3.605896 loss_obj: 4.028077 loss_cls: 1.084714 loss: 10.701500 eta: 2:31:45 batch_cost: 0.9889 data_cost: 0.6041 ips: 32.3583 images/s
[04/14 13:52:34] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:52:37] ppdet.engine INFO: Epoch: [75] [ 0/18] learning_rate: 0.001688 loss_xy: 0.947432 loss_wh: 0.910381 loss_iou: 3.634017 loss_obj: 4.010726 loss_cls: 1.041002 loss: 10.577214 eta: 2:31:02 batch_cost: 0.8199 data_cost: 0.4567 ips: 39.0311 images/s
[04/14 13:52:55] ppdet.engine INFO: Epoch: [76] [ 0/18] learning_rate: 0.001710 loss_xy: 0.955729 loss_wh: 0.912969 loss_iou: 3.629117 loss_obj: 4.104681 loss_cls: 1.045977 loss: 10.479834 eta: 2:30:44 batch_cost: 0.9333 data_cost: 0.5302 ips: 34.2856 images/s
[04/14 13:53:13] ppdet.engine INFO: Epoch: [77] [ 0/18] learning_rate: 0.001732 loss_xy: 0.974595 loss_wh: 0.963698 loss_iou: 3.676403 loss_obj: 4.206381 loss_cls: 1.012113 loss: 10.845070 eta: 2:30:20 batch_cost: 0.9067 data_cost: 0.5173 ips: 35.2939 images/s
[04/14 13:53:30] ppdet.engine INFO: Epoch: [78] [ 0/18] learning_rate: 0.001755 loss_xy: 0.967862 loss_wh: 0.909415 loss_iou: 3.533006 loss_obj: 4.054981 loss_cls: 1.030352 loss: 10.537533 eta: 2:29:51 batch_cost: 0.8964 data_cost: 0.4785 ips: 35.7003 images/s
[04/14 13:53:52] ppdet.engine INFO: Epoch: [79] [ 0/18] learning_rate: 0.001778 loss_xy: 0.975711 loss_wh: 0.914810 loss_iou: 3.644373 loss_obj: 4.328164 loss_cls: 1.048790 loss: 11.067169 eta: 2:29:56 batch_cost: 1.1354 data_cost: 0.7344 ips: 28.1838 images/s
[04/14 13:54:08] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:54:11] ppdet.engine INFO: Epoch: [80] [ 0/18] learning_rate: 0.001800 loss_xy: 0.974182 loss_wh: 0.967767 loss_iou: 3.771420 loss_obj: 4.259057 loss_cls: 1.064002 loss: 11.044782 eta: 2:29:39 batch_cost: 1.0820 data_cost: 0.6572 ips: 29.5753 images/s
[04/14 13:54:31] ppdet.engine INFO: Epoch: [81] [ 0/18] learning_rate: 0.001822 loss_xy: 0.984474 loss_wh: 1.069915 loss_iou: 3.966429 loss_obj: 4.724171 loss_cls: 1.035542 loss: 11.411337 eta: 2:29:29 batch_cost: 1.0230 data_cost: 0.6373 ips: 31.2816 images/s
[04/14 13:54:48] ppdet.engine INFO: Epoch: [82] [ 0/18] learning_rate: 0.001845 loss_xy: 0.987013 loss_wh: 0.945098 loss_iou: 3.455522 loss_obj: 4.319368 loss_cls: 0.951071 loss: 10.577715 eta: 2:29:09 batch_cost: 0.9729 data_cost: 0.5607 ips: 32.8921 images/s
[04/14 13:55:08] ppdet.engine INFO: Epoch: [83] [ 0/18] learning_rate: 0.001868 loss_xy: 0.979008 loss_wh: 0.958179 loss_iou: 3.594563 loss_obj: 4.100698 loss_cls: 0.976556 loss: 10.620663 eta: 2:29:02 batch_cost: 1.0225 data_cost: 0.6501 ips: 31.2971 images/s
[04/14 13:55:27] ppdet.engine INFO: Epoch: [84] [ 0/18] learning_rate: 0.001890 loss_xy: 0.949782 loss_wh: 0.893462 loss_iou: 3.698454 loss_obj: 4.125708 loss_cls: 1.005782 loss: 10.740036 eta: 2:28:48 batch_cost: 1.0790 data_cost: 0.6847 ips: 29.6567 images/s
[04/14 13:55:42] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:55:45] ppdet.engine INFO: Epoch: [85] [ 0/18] learning_rate: 0.001913 loss_xy: 0.943198 loss_wh: 0.849851 loss_iou: 3.307967 loss_obj: 3.922647 loss_cls: 0.944305 loss: 10.135038 eta: 2:28:24 batch_cost: 0.9472 data_cost: 0.5616 ips: 33.7845 images/s
[04/14 13:56:06] ppdet.engine INFO: Epoch: [86] [ 0/18] learning_rate: 0.001935 loss_xy: 0.925228 loss_wh: 0.868589 loss_iou: 3.516091 loss_obj: 3.981165 loss_cls: 0.903726 loss: 10.259991 eta: 2:28:19 batch_cost: 1.0563 data_cost: 0.6453 ips: 30.2952 images/s
[04/14 13:56:24] ppdet.engine INFO: Epoch: [87] [ 0/18] learning_rate: 0.001958 loss_xy: 0.975347 loss_wh: 0.859342 loss_iou: 3.439246 loss_obj: 4.020237 loss_cls: 0.862954 loss: 10.102800 eta: 2:28:02 batch_cost: 1.1076 data_cost: 0.7201 ips: 28.8903 images/s
[04/14 13:56:44] ppdet.engine INFO: Epoch: [88] [ 0/18] learning_rate: 0.001980 loss_xy: 0.960975 loss_wh: 0.885990 loss_iou: 3.448644 loss_obj: 3.971878 loss_cls: 0.871910 loss: 10.308729 eta: 2:27:55 batch_cost: 1.1498 data_cost: 0.7033 ips: 27.8304 images/s
[04/14 13:57:03] ppdet.engine INFO: Epoch: [89] [ 0/18] learning_rate: 0.002002 loss_xy: 0.952255 loss_wh: 0.943906 loss_iou: 3.452410 loss_obj: 4.157928 loss_cls: 0.812853 loss: 10.337480 eta: 2:27:38 batch_cost: 1.0300 data_cost: 0.6007 ips: 31.0689 images/s
[04/14 13:57:20] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:57:23] ppdet.engine INFO: Epoch: [90] [ 0/18] learning_rate: 0.002025 loss_xy: 0.950907 loss_wh: 0.970275 loss_iou: 3.525810 loss_obj: 4.221418 loss_cls: 0.941517 loss: 10.438553 eta: 2:27:30 batch_cost: 1.0413 data_cost: 0.6033 ips: 30.7304 images/s
[04/14 13:57:46] ppdet.engine INFO: Epoch: [91] [ 0/18] learning_rate: 0.002047 loss_xy: 0.996558 loss_wh: 0.978084 loss_iou: 3.572124 loss_obj: 4.101471 loss_cls: 0.829481 loss: 10.548926 eta: 2:27:37 batch_cost: 1.1513 data_cost: 0.7074 ips: 27.7935 images/s
[04/14 13:58:06] ppdet.engine INFO: Epoch: [92] [ 0/18] learning_rate: 0.002070 loss_xy: 0.970499 loss_wh: 0.926036 loss_iou: 3.523901 loss_obj: 3.947544 loss_cls: 0.845980 loss: 10.222193 eta: 2:27:29 batch_cost: 1.0978 data_cost: 0.6706 ips: 29.1488 images/s
[04/14 13:58:25] ppdet.engine INFO: Epoch: [93] [ 0/18] learning_rate: 0.002093 loss_xy: 0.959820 loss_wh: 0.905338 loss_iou: 3.550429 loss_obj: 4.114080 loss_cls: 0.855164 loss: 10.316627 eta: 2:27:18 batch_cost: 1.1143 data_cost: 0.7821 ips: 28.7171 images/s
[04/14 13:58:48] ppdet.engine INFO: Epoch: [94] [ 0/18] learning_rate: 0.002115 loss_xy: 0.981444 loss_wh: 0.888150 loss_iou: 3.415428 loss_obj: 4.109206 loss_cls: 0.798827 loss: 10.335572 eta: 2:27:26 batch_cost: 1.2525 data_cost: 0.8564 ips: 25.5489 images/s
[04/14 13:59:05] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 13:59:10] ppdet.engine INFO: Epoch: [95] [ 0/18] learning_rate: 0.002138 loss_xy: 0.955837 loss_wh: 0.873673 loss_iou: 3.332043 loss_obj: 3.809483 loss_cls: 0.751365 loss: 9.813274 eta: 2:27:23 batch_cost: 1.2752 data_cost: 0.8334 ips: 25.0936 images/s
[04/14 13:59:31] ppdet.engine INFO: Epoch: [96] [ 0/18] learning_rate: 0.002160 loss_xy: 0.935621 loss_wh: 0.852627 loss_iou: 3.414521 loss_obj: 3.812283 loss_cls: 0.802028 loss: 10.158546 eta: 2:27:22 batch_cost: 1.1934 data_cost: 0.7717 ips: 26.8142 images/s
[04/14 13:59:48] ppdet.engine INFO: Epoch: [97] [ 0/18] learning_rate: 0.002182 loss_xy: 0.946182 loss_wh: 0.847514 loss_iou: 3.410558 loss_obj: 3.885823 loss_cls: 0.753103 loss: 9.827497 eta: 2:26:59 batch_cost: 1.0255 data_cost: 0.6493 ips: 31.2043 images/s
[04/14 14:00:09] ppdet.engine INFO: Epoch: [98] [ 0/18] learning_rate: 0.002205 loss_xy: 0.943210 loss_wh: 0.890625 loss_iou: 3.494732 loss_obj: 3.937510 loss_cls: 0.862669 loss: 10.151651 eta: 2:26:50 batch_cost: 1.0532 data_cost: 0.6395 ips: 30.3841 images/s
[04/14 14:00:29] ppdet.engine INFO: Epoch: [99] [ 0/18] learning_rate: 0.002227 loss_xy: 0.938594 loss_wh: 0.899218 loss_iou: 3.540349 loss_obj: 3.690251 loss_cls: 0.845537 loss: 9.925249 eta: 2:26:42 batch_cost: 1.0790 data_cost: 0.6809 ips: 29.6581 images/s
[04/14 14:00:44] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 14:00:47] ppdet.engine INFO: Epoch: [100] [ 0/18] learning_rate: 0.002250 loss_xy: 0.919214 loss_wh: 0.916093 loss_iou: 3.426758 loss_obj: 3.755184 loss_cls: 0.764831 loss: 10.005158 eta: 2:26:19 batch_cost: 0.9922 data_cost: 0.5528 ips: 32.2505 images/s
[04/14 14:01:06] ppdet.engine INFO: Epoch: [101] [ 0/18] learning_rate: 0.002273 loss_xy: 0.937125 loss_wh: 0.900280 loss_iou: 3.460453 loss_obj: 3.921981 loss_cls: 0.705801 loss: 10.027704 eta: 2:26:05 batch_cost: 1.0060 data_cost: 0.5976 ips: 31.8079 images/s
[04/14 14:01:29] ppdet.engine INFO: Epoch: [102] [ 0/18] learning_rate: 0.002295 loss_xy: 0.929607 loss_wh: 0.920467 loss_iou: 3.616426 loss_obj: 3.982140 loss_cls: 0.777069 loss: 10.429827 eta: 2:26:07 batch_cost: 1.1968 data_cost: 0.7357 ips: 26.7383 images/s
[04/14 14:01:45] ppdet.engine INFO: Epoch: [103] [ 0/18] learning_rate: 0.002318 loss_xy: 0.887020 loss_wh: 0.826823 loss_iou: 3.299038 loss_obj: 3.873783 loss_cls: 0.739317 loss: 9.589198 eta: 2:25:36 batch_cost: 0.9801 data_cost: 0.5898 ips: 32.6511 images/s
[04/14 14:02:03] ppdet.engine INFO: Epoch: [104] [ 0/18] learning_rate: 0.002340 loss_xy: 0.965848 loss_wh: 0.879198 loss_iou: 3.397014 loss_obj: 3.819821 loss_cls: 0.751390 loss: 9.773285 eta: 2:25:19 batch_cost: 0.9593 data_cost: 0.5575 ips: 33.3562 images/s
[04/14 14:02:19] ppdet.utils.checkpoint INFO: Save checkpoint: output/model
[04/14 14:02:24] ppdet.engine INFO: Epoch: [105] [ 0/18] learning_rate: 0.002363 loss_xy: 0.898395 loss_wh: 0.948094 loss_iou: 3.442998 loss_obj: 3.933207 loss_cls: 0.748045 loss: 10.250551 eta: 2:25:12 batch_cost: 1.0266 data_cost: 0.6346 ips: 31.1720 images/s
[04/14 14:02:45] ppdet.engine INFO: Epoch: [106] [ 0/18] learning_rate: 0.002385 loss_xy: 0.962784 loss_wh: 1.034158 loss_iou: 3.827223 loss_obj: 4.143725 loss_cls: 0.701851 loss: 10.612465 eta: 2:25:03 batch_cost: 1.1265 data_cost: 0.6973 ips: 28.4062 images/s
[04/14 14:03:03] ppdet.engine INFO: Epoch: [107] [ 0/18] learning_rate: 0.002407 loss_xy: 0.941749 loss_wh: 0.944827 loss_iou: 3.561562 loss_obj: 4.014892 loss_cls: 0.688428 loss: 10.248665 eta: 2:24:45 batch_cost: 1.0110 data_cost: 0.6108 ips: 31.6509 images/s
[04/14 14:03:19] ppdet.engine INFO: Epoch: [108] [ 0/18] learning_rate: 0.002430 loss_xy: 0.926247 loss_wh: 0.856184 loss_iou: 3.391245 loss_obj: 3.796605 loss_cls: 0.705016 loss: 9.728722 eta: 2:24:19 batch_cost: 0.9577 data_cost: 0.5569 ips: 33.4120 images/s
^C
Traceback (most recent call last):
File "tools/train.py", line 177, in <module>
main()
File "tools/train.py", line 173, in main
run(FLAGS, cfg)
File "tools/train.py", line 127, in run
trainer.train(FLAGS.eval)
File "/home/aistudio/PaddleDetection/ppdet/engine/trainer.py", line 445, in train
self.ema.update()
File "/home/aistudio/PaddleDetection/ppdet/optimizer.py", line 387, in update
v = decay * v + (1 - decay) * model_dict[k]
File "/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/math_op_patch.py", line 264, in __impl__
return math_op(self, other_var, 'axis', axis)
KeyboardInterrupt
In [24]
# 查看推理效果
import os
os.chdir('PaddleDetection')
!export CUDA_VISIBLE_DEVICES=0 #windows和Mac下不需要执行该命令
!python tools/infer.py -c configs/ppyolo/model.yml \
--infer_img=/home/aistudio/data/images/1.jpg \
--output_dir=infer_output/ \
--draw_threshold=0.5 \
-o weights=output/iter_100/69.pdparams \
--use_vdl=Ture
Warning: import ppdet from source directory without installing, run 'python setup.py install' to install ppdet firstly W0414 14:06:48.324729 3731 device_context.cc:447] Please NOTE: device: 0, GPU Compute Capability: 7.0, Driver API Version: 10.1, Runtime API Version: 10.1 W0414 14:06:48.330070 3731 device_context.cc:465] device: 0, cuDNN Version: 7.6. [04/14 14:06:51] ppdet.utils.checkpoint INFO: ['backbone.conv1._batch_norm._mean', 'backbone.conv1._batch_norm._variance', 'backbone.conv1._batch_norm.bias', 'backbone.conv1._batch_norm.weight', 'backbone.conv1._conv.weight', 'backbone.conv2_1._depthwise_conv._batch_norm._mean', 'backbone.conv2_1._depthwise_conv._batch_norm._variance', 'backbone.conv2_1._depthwise_conv._batch_norm.bias', 'backbone.conv2_1._depthwise_conv._batch_norm.weight', 'backbone.conv2_1._depthwise_conv._conv.weight', 'backbone.conv2_1._pointwise_conv._batch_norm._mean', 'backbone.conv2_1._pointwise_conv._batch_norm._variance', 'backbone.conv2_1._pointwise_conv._batch_norm.bias', 'backbone.conv2_1._pointwise_conv._batch_norm.weight', 'backbone.conv2_1._pointwise_conv._conv.weight', 'backbone.conv2_2._depthwise_conv._batch_norm._mean', 'backbone.conv2_2._depthwise_conv._batch_norm._variance', 'backbone.conv2_2._depthwise_conv._batch_norm.bias', 'backbone.conv2_2._depthwise_conv._batch_norm.weight', 'backbone.conv2_2._depthwise_conv._conv.weight', 'backbone.conv2_2._pointwise_conv._batch_norm._mean', 'backbone.conv2_2._pointwise_conv._batch_norm._variance', 'backbone.conv2_2._pointwise_conv._batch_norm.bias', 'backbone.conv2_2._pointwise_conv._batch_norm.weight', 'backbone.conv2_2._pointwise_conv._conv.weight', 'backbone.conv3_1._depthwise_conv._batch_norm._mean', 'backbone.conv3_1._depthwise_conv._batch_norm._variance', 'backbone.conv3_1._depthwise_conv._batch_norm.bias', 'backbone.conv3_1._depthwise_conv._batch_norm.weight', 'backbone.conv3_1._depthwise_conv._conv.weight', 'backbone.conv3_1._pointwise_conv._batch_norm._mean', 'backbone.conv3_1._pointwise_conv._batch_norm._variance', 'backbone.conv3_1._pointwise_conv._batch_norm.bias', 'backbone.conv3_1._pointwise_conv._batch_norm.weight', 'backbone.conv3_1._pointwise_conv._conv.weight', 'backbone.conv3_2._depthwise_conv._batch_norm._mean', 'backbone.conv3_2._depthwise_conv._batch_norm._variance', 'backbone.conv3_2._depthwise_conv._batch_norm.bias', 'backbone.conv3_2._depthwise_conv._batch_norm.weight', 'backbone.conv3_2._depthwise_conv._conv.weight', 'backbone.conv3_2._pointwise_conv._batch_norm._mean', 'backbone.conv3_2._pointwise_conv._batch_norm._variance', 'backbone.conv3_2._pointwise_conv._batch_norm.bias', 'backbone.conv3_2._pointwise_conv._batch_norm.weight', 'backbone.conv3_2._pointwise_conv._conv.weight', 'backbone.conv4_1._depthwise_conv._batch_norm._mean', 'backbone.conv4_1._depthwise_conv._batch_norm._variance', 'backbone.conv4_1._depthwise_conv._batch_norm.bias', 'backbone.conv4_1._depthwise_conv._batch_norm.weight', 'backbone.conv4_1._depthwise_conv._conv.weight', 'backbone.conv4_1._pointwise_conv._batch_norm._mean', 'backbone.conv4_1._pointwise_conv._batch_norm._variance', 'backbone.conv4_1._pointwise_conv._batch_norm.bias', 'backbone.conv4_1._pointwise_conv._batch_norm.weight', 'backbone.conv4_1._pointwise_conv._conv.weight', 'backbone.conv4_2._depthwise_conv._batch_norm._mean', 'backbone.conv4_2._depthwise_conv._batch_norm._variance', 'backbone.conv4_2._depthwise_conv._batch_norm.bias', 'backbone.conv4_2._depthwise_conv._batch_norm.weight', 'backbone.conv4_2._depthwise_conv._conv.weight', 'backbone.conv4_2._pointwise_conv._batch_norm._mean', 'backbone.conv4_2._pointwise_conv._batch_norm._variance', 'backbone.conv4_2._pointwise_conv._batch_norm.bias', 'backbone.conv4_2._pointwise_conv._batch_norm.weight', 'backbone.conv4_2._pointwise_conv._conv.weight', 'backbone.conv5_1._depthwise_conv._batch_norm._mean', 'backbone.conv5_1._depthwise_conv._batch_norm._variance', 'backbone.conv5_1._depthwise_conv._batch_norm.bias', 'backbone.conv5_1._depthwise_conv._batch_norm.weight', 'backbone.conv5_1._depthwise_conv._conv.weight', 'backbone.conv5_1._pointwise_conv._batch_norm._mean', 'backbone.conv5_1._pointwise_conv._batch_norm._variance', 'backbone.conv5_1._pointwise_conv._batch_norm.bias', 'backbone.conv5_1._pointwise_conv._batch_norm.weight', 'backbone.conv5_1._pointwise_conv._conv.weight', 'backbone.conv5_2._depthwise_conv._batch_norm._mean', 'backbone.conv5_2._depthwise_conv._batch_norm._variance', 'backbone.conv5_2._depthwise_conv._batch_norm.bias', 'backbone.conv5_2._depthwise_conv._batch_norm.weight', 'backbone.conv5_2._depthwise_conv._conv.weight', 'backbone.conv5_2._pointwise_conv._batch_norm._mean', 'backbone.conv5_2._pointwise_conv._batch_norm._variance', 'backbone.conv5_2._pointwise_conv._batch_norm.bias', 'backbone.conv5_2._pointwise_conv._batch_norm.weight', 'backbone.conv5_2._pointwise_conv._conv.weight', 'backbone.conv5_3._depthwise_conv._batch_norm._mean', 'backbone.conv5_3._depthwise_conv._batch_norm._variance', 'backbone.conv5_3._depthwise_conv._batch_norm.bias', 'backbone.conv5_3._depthwise_conv._batch_norm.weight', 'backbone.conv5_3._depthwise_conv._conv.weight', 'backbone.conv5_3._pointwise_conv._batch_norm._mean', 'backbone.conv5_3._pointwise_conv._batch_norm._variance', 'backbone.conv5_3._pointwise_conv._batch_norm.bias', 'backbone.conv5_3._pointwise_conv._batch_norm.weight', 'backbone.conv5_3._pointwise_conv._conv.weight', 'backbone.conv5_4._depthwise_conv._batch_norm._mean', 'backbone.conv5_4._depthwise_conv._batch_norm._variance', 'backbone.conv5_4._depthwise_conv._batch_norm.bias', 'backbone.conv5_4._depthwise_conv._batch_norm.weight', 'backbone.conv5_4._depthwise_conv._conv.weight', 'backbone.conv5_4._pointwise_conv._batch_norm._mean', 'backbone.conv5_4._pointwise_conv._batch_norm._variance', 'backbone.conv5_4._pointwise_conv._batch_norm.bias', 'backbone.conv5_4._pointwise_conv._batch_norm.weight', 'backbone.conv5_4._pointwise_conv._conv.weight', 'backbone.conv5_5._depthwise_conv._batch_norm._mean', 'backbone.conv5_5._depthwise_conv._batch_norm._variance', 'backbone.conv5_5._depthwise_conv._batch_norm.bias', 'backbone.conv5_5._depthwise_conv._batch_norm.weight', 'backbone.conv5_5._depthwise_conv._conv.weight', 'backbone.conv5_5._pointwise_conv._batch_norm._mean', 'backbone.conv5_5._pointwise_conv._batch_norm._variance', 'backbone.conv5_5._pointwise_conv._batch_norm.bias', 'backbone.conv5_5._pointwise_conv._batch_norm.weight', 'backbone.conv5_5._pointwise_conv._conv.weight', 'backbone.conv5_6._depthwise_conv._batch_norm._mean', 'backbone.conv5_6._depthwise_conv._batch_norm._variance', 'backbone.conv5_6._depthwise_conv._batch_norm.bias', 'backbone.conv5_6._depthwise_conv._batch_norm.weight', 'backbone.conv5_6._depthwise_conv._conv.weight', 'backbone.conv5_6._pointwise_conv._batch_norm._mean', 'backbone.conv5_6._pointwise_conv._batch_norm._variance', 'backbone.conv5_6._pointwise_conv._batch_norm.bias', 'backbone.conv5_6._pointwise_conv._batch_norm.weight', 'backbone.conv5_6._pointwise_conv._conv.weight', 'backbone.conv6._depthwise_conv._batch_norm._mean', 'backbone.conv6._depthwise_conv._batch_norm._variance', 'backbone.conv6._depthwise_conv._batch_norm.bias', 'backbone.conv6._depthwise_conv._batch_norm.weight', 'backbone.conv6._depthwise_conv._conv.weight', 'backbone.conv6._pointwise_conv._batch_norm._mean', 'backbone.conv6._pointwise_conv._batch_norm._variance', 'backbone.conv6._pointwise_conv._batch_norm.bias', 'backbone.conv6._pointwise_conv._batch_norm.weight', 'backbone.conv6._pointwise_conv._conv.weight', 'neck.yolo_block.0.conv_module.conv0.batch_norm._mean', 'neck.yolo_block.0.conv_module.conv0.batch_norm._variance', 'neck.yolo_block.0.conv_module.conv0.batch_norm.bias', 'neck.yolo_block.0.conv_module.conv0.batch_norm.weight', 'neck.yolo_block.0.conv_module.conv0.conv.weight', 'neck.yolo_block.0.conv_module.conv1.batch_norm._mean', 'neck.yolo_block.0.conv_module.conv1.batch_norm._variance', 'neck.yolo_block.0.conv_module.conv1.batch_norm.bias', 'neck.yolo_block.0.conv_module.conv1.batch_norm.weight', 'neck.yolo_block.0.conv_module.conv1.conv.weight', 'neck.yolo_block.0.conv_module.conv2.batch_norm._mean', 'neck.yolo_block.0.conv_module.conv2.batch_norm._variance', 'neck.yolo_block.0.conv_module.conv2.batch_norm.bias', 'neck.yolo_block.0.conv_module.conv2.batch_norm.weight', 'neck.yolo_block.0.conv_module.conv2.conv.weight', 'neck.yolo_block.0.conv_module.conv3.batch_norm._mean', 'neck.yolo_block.0.conv_module.conv3.batch_norm._variance', 'neck.yolo_block.0.conv_module.conv3.batch_norm.bias', 'neck.yolo_block.0.conv_module.conv3.batch_norm.weight', 'neck.yolo_block.0.conv_module.conv3.conv.weight', 'neck.yolo_block.0.conv_module.route.batch_norm._mean', 'neck.yolo_block.0.conv_module.route.batch_norm._variance', 'neck.yolo_block.0.conv_module.route.batch_norm.bias', 'neck.yolo_block.0.conv_module.route.batch_norm.weight', 'neck.yolo_block.0.conv_module.route.conv.weight', 'neck.yolo_block.1.conv_module.conv0.batch_norm._mean', 'neck.yolo_block.1.conv_module.conv0.batch_norm._variance', 'neck.yolo_block.1.conv_module.conv0.batch_norm.bias', 'neck.yolo_block.1.conv_module.conv0.batch_norm.weight', 'neck.yolo_block.1.conv_module.conv0.conv.weight', 'neck.yolo_block.1.conv_module.conv1.batch_norm._mean', 'neck.yolo_block.1.conv_module.conv1.batch_norm._variance', 'neck.yolo_block.1.conv_module.conv1.batch_norm.bias', 'neck.yolo_block.1.conv_module.conv1.batch_norm.weight', 'neck.yolo_block.1.conv_module.conv1.conv.weight', 'neck.yolo_block.1.conv_module.conv2.batch_norm._mean', 'neck.yolo_block.1.conv_module.conv2.batch_norm._variance', 'neck.yolo_block.1.conv_module.conv2.batch_norm.bias', 'neck.yolo_block.1.conv_module.conv2.batch_norm.weight', 'neck.yolo_block.1.conv_module.conv2.conv.weight', 'neck.yolo_block.1.conv_module.conv3.batch_norm._mean', 'neck.yolo_block.1.conv_module.conv3.batch_norm._variance', 'neck.yolo_block.1.conv_module.conv3.batch_norm.bias', 'neck.yolo_block.1.conv_module.conv3.batch_norm.weight', 'neck.yolo_block.1.conv_module.conv3.conv.weight', 'neck.yolo_block.1.conv_module.route.batch_norm._mean', 'neck.yolo_block.1.conv_module.route.batch_norm._variance', 'neck.yolo_block.1.conv_module.route.batch_norm.bias', 'neck.yolo_block.1.conv_module.route.batch_norm.weight', 'neck.yolo_block.1.conv_module.route.conv.weight', 'neck.yolo_block.2.conv_module.conv0.batch_norm._mean', 'neck.yolo_block.2.conv_module.conv0.batch_norm._variance', 'neck.yolo_block.2.conv_module.conv0.batch_norm.bias', 'neck.yolo_block.2.conv_module.conv0.batch_norm.weight', 'neck.yolo_block.2.conv_module.conv0.conv.weight', 'neck.yolo_block.2.conv_module.conv1.batch_norm._mean', 'neck.yolo_block.2.conv_module.conv1.batch_norm._variance', 'neck.yolo_block.2.conv_module.conv1.batch_norm.bias', 'neck.yolo_block.2.conv_module.conv1.batch_norm.weight', 'neck.yolo_block.2.conv_module.conv1.conv.weight', 'neck.yolo_block.2.conv_module.conv2.batch_norm._mean', 'neck.yolo_block.2.conv_module.conv2.batch_norm._variance', 'neck.yolo_block.2.conv_module.conv2.batch_norm.bias', 'neck.yolo_block.2.conv_module.conv2.batch_norm.weight', 'neck.yolo_block.2.conv_module.conv2.conv.weight', 'neck.yolo_block.2.conv_module.conv3.batch_norm._mean', 'neck.yolo_block.2.conv_module.conv3.batch_norm._variance', 'neck.yolo_block.2.conv_module.conv3.batch_norm.bias', 'neck.yolo_block.2.conv_module.conv3.batch_norm.weight', 'neck.yolo_block.2.conv_module.conv3.conv.weight', 'neck.yolo_block.2.conv_module.route.batch_norm._mean', 'neck.yolo_block.2.conv_module.route.batch_norm._variance', 'neck.yolo_block.2.conv_module.route.batch_norm.bias', 'neck.yolo_block.2.conv_module.route.batch_norm.weight', 'neck.yolo_block.2.conv_module.route.conv.weight'] in pretrained weight is not used in the model, and its will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [1024] in pretrained weight neck.yolo_block.0.tip.batch_norm._mean is unmatched with the shape [160] in model neck.yolo_block.0.tip.batch_norm._mean. And the weight neck.yolo_block.0.tip.batch_norm._mean will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [1024] in pretrained weight neck.yolo_block.0.tip.batch_norm._variance is unmatched with the shape [160] in model neck.yolo_block.0.tip.batch_norm._variance. And the weight neck.yolo_block.0.tip.batch_norm._variance will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [1024] in pretrained weight neck.yolo_block.0.tip.batch_norm.bias is unmatched with the shape [160] in model neck.yolo_block.0.tip.batch_norm.bias. And the weight neck.yolo_block.0.tip.batch_norm.bias will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [1024] in pretrained weight neck.yolo_block.0.tip.batch_norm.weight is unmatched with the shape [160] in model neck.yolo_block.0.tip.batch_norm.weight. And the weight neck.yolo_block.0.tip.batch_norm.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [1024, 512, 3, 3] in pretrained weight neck.yolo_block.0.tip.conv.weight is unmatched with the shape [160, 160, 1, 1] in model neck.yolo_block.0.tip.conv.weight. And the weight neck.yolo_block.0.tip.conv.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [512] in pretrained weight neck.yolo_block.1.tip.batch_norm._mean is unmatched with the shape [128] in model neck.yolo_block.1.tip.batch_norm._mean. And the weight neck.yolo_block.1.tip.batch_norm._mean will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [512] in pretrained weight neck.yolo_block.1.tip.batch_norm._variance is unmatched with the shape [128] in model neck.yolo_block.1.tip.batch_norm._variance. And the weight neck.yolo_block.1.tip.batch_norm._variance will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [512] in pretrained weight neck.yolo_block.1.tip.batch_norm.bias is unmatched with the shape [128] in model neck.yolo_block.1.tip.batch_norm.bias. And the weight neck.yolo_block.1.tip.batch_norm.bias will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [512] in pretrained weight neck.yolo_block.1.tip.batch_norm.weight is unmatched with the shape [128] in model neck.yolo_block.1.tip.batch_norm.weight. And the weight neck.yolo_block.1.tip.batch_norm.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [512, 256, 3, 3] in pretrained weight neck.yolo_block.1.tip.conv.weight is unmatched with the shape [128, 128, 1, 1] in model neck.yolo_block.1.tip.conv.weight. And the weight neck.yolo_block.1.tip.conv.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_block.2.tip.batch_norm._mean is unmatched with the shape [96] in model neck.yolo_block.2.tip.batch_norm._mean. And the weight neck.yolo_block.2.tip.batch_norm._mean will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_block.2.tip.batch_norm._variance is unmatched with the shape [96] in model neck.yolo_block.2.tip.batch_norm._variance. And the weight neck.yolo_block.2.tip.batch_norm._variance will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_block.2.tip.batch_norm.bias is unmatched with the shape [96] in model neck.yolo_block.2.tip.batch_norm.bias. And the weight neck.yolo_block.2.tip.batch_norm.bias will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_block.2.tip.batch_norm.weight is unmatched with the shape [96] in model neck.yolo_block.2.tip.batch_norm.weight. And the weight neck.yolo_block.2.tip.batch_norm.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256, 128, 3, 3] in pretrained weight neck.yolo_block.2.tip.conv.weight is unmatched with the shape [96, 96, 1, 1] in model neck.yolo_block.2.tip.conv.weight. And the weight neck.yolo_block.2.tip.conv.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_transition.0.batch_norm._mean is unmatched with the shape [160] in model neck.yolo_transition.0.batch_norm._mean. And the weight neck.yolo_transition.0.batch_norm._mean will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_transition.0.batch_norm._variance is unmatched with the shape [160] in model neck.yolo_transition.0.batch_norm._variance. And the weight neck.yolo_transition.0.batch_norm._variance will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_transition.0.batch_norm.bias is unmatched with the shape [160] in model neck.yolo_transition.0.batch_norm.bias. And the weight neck.yolo_transition.0.batch_norm.bias will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256] in pretrained weight neck.yolo_transition.0.batch_norm.weight is unmatched with the shape [160] in model neck.yolo_transition.0.batch_norm.weight. And the weight neck.yolo_transition.0.batch_norm.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [256, 512, 1, 1] in pretrained weight neck.yolo_transition.0.conv.weight is unmatched with the shape [160, 160, 1, 1] in model neck.yolo_transition.0.conv.weight. And the weight neck.yolo_transition.0.conv.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [128, 256, 1, 1] in pretrained weight neck.yolo_transition.1.conv.weight is unmatched with the shape [128, 128, 1, 1] in model neck.yolo_transition.1.conv.weight. And the weight neck.yolo_transition.1.conv.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [21, 1024, 1, 1] in pretrained weight yolo_head.yolo_output.0.weight is unmatched with the shape [21, 160, 1, 1] in model yolo_head.yolo_output.0.weight. And the weight yolo_head.yolo_output.0.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [21, 512, 1, 1] in pretrained weight yolo_head.yolo_output.1.weight is unmatched with the shape [21, 128, 1, 1] in model yolo_head.yolo_output.1.weight. And the weight yolo_head.yolo_output.1.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: The shape [21, 256, 1, 1] in pretrained weight yolo_head.yolo_output.2.weight is unmatched with the shape [21, 96, 1, 1] in model yolo_head.yolo_output.2.weight. And the weight yolo_head.yolo_output.2.weight will not be loaded [04/14 14:06:51] ppdet.utils.checkpoint INFO: Finish loading model weights: output/iter_100/69.pdparams loading annotations into memory... Done (t=0.00s) creating index... index created! 100%|█████████████████████████████████████████████| 1/1 [00:00<00:00, 5.03it/s] [04/14 14:06:51] ppdet.engine INFO: Detection bbox results save in infer_output/1.jpg
4. 转化为静态图
.pdiparams、 .pdmodel、 .pdiparams.info、 infer_cfg.yml
In [29]
os.chdir('PaddleDetection')
!python tools/export_model.py -c configs/ppyolo/model.yml \
--output_dir=./inference_model \
-o weights=output/model/94
os.chdir('/home/aistudio')^C
In [28]
os.chdir('/home/aistudio')
!zip -r -q -o model.zip PaddleDetection/inference_model/model5.在JetsonNano部署
参考https://aistudio.baidu.com/aistudio/projectdetail/3451173
5.1 nano环境
- jetpack 4.4
- tensorrt 0.7.0
- paddle 2.2.2 https://paddleinference.paddlepaddle.org.cn/user_guides/download_lib.html#python
5.2 infer.py
import cv2
import numpy as np
from paddle.inference import Config
from paddle.inference import PrecisionType
from paddle.inference import create_predictor
import yaml
import time
def resize(img, target_size):
"resize to target_size"
if not isinstance(img,np.ndarray):
raise TypeError('Image type is not numpy.')
im_shape = img.shape
im_size_min = np.min(im_shape[0:2])
im_size_max = np.max(im_shape[0:2])
im_scale_x = float(target_size) / float(im_shape[1])
im_scale_y = float(target_size) / float(im_shape[0])
img = cv2.resize(img, None, None, fx=im_scale_x, fy=im_scale_y)
return img
def normalize(img, mean, std):
img = img / 255.0
mean = np.array(mean)[np.newaxis, np.newaxis, :]
std = np.array(std)[np.newaxis, np.newaxis, :]
img -= mean
img /= std
return img
def preprocess(img,img_size):
mean = [0.485, 0.456, 0.406]
std = [0.229, 0.224, 0.225]
img = resize(img, img_size)
img = img[:, :, ::-1].astype('float32')
img = normalize(img, mean, std)
img = img.transpose((2, 0, 1))
return img[np.newaxis, :]
def predict_config(model_file, params_file):
config = Config()
config.set_prog_file(model_file)
config.set_params_file(params_file)
# Config's default setting is CPU, we should open GPU manully
config.enable_use_gpu(500, 0)
config.switch_ir_optim()
config.enable_memory_optim()
config.enable_tensorrt_engine(workspace_size=1 << 30, precision_mode = PrecisionType.Float32, max_batch_size=1, min_subgraph_size=5, use_static=False, use_calib_mode=False)
predictor = create_predictor(config)
return predictor
def predict(predictor, img):
'''
fun: init predictor
input: (.pdmodel,.pdiparams)
output: class predictor
'''
input_names = predictor.get_input_names()
for i, name in enumerate(input_names):
input_tensor = predictor.get_input_handle(name)
input_tensor.reshape(img[i].shape)
input_tensor.copy_from_cpu(img[i].copy())
# execute predictor
predictor.run()
# get output
results = []
output_names = predictor.get_output_names()
for i, name in enumerate(output_names):
output_tensor = predictor.get_output_handle(name)
output_data = output_tensor.copy_to_cpu()
results.append(output_data)
return results
def draw_bbox_image(frame, result, label_list, fps, threshold=0.5):
for res in result:
cat_id, score, bbox = res[0], res[1], res[2:]
if score < threshold:
continue
for i in bbox:
int(i)
xmin, ymin, xmax, ymax = bbox
cv2.rectangle(frame, (xmin, ymin), (xmax, ymax), (255, 0, 255), 2)
print('category id is :{}, bbox is :{}'.format(cat_id, bbox))
try:
label_id = label_list[int(cat_id)]
cv2.putText(frame,'FPS:'+str(fps) , (20,20),cv2.FONT_HERSHEY_SIMPLEX, 0.5,(255,0,0), 2)
cv2.putText(frame, label_id, (int(xmin),int(ymin-2)), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (255,0,0), 2)
cv2.putText(frame, str(round(score,2)), (int(xmin-35),int(ymin-2)), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0,255,0), 2)
except KeyError:
pass
# MAIN
def main():
infer_cfg = open('inference_model/model/infer_cfg.yml')
data = infer_cfg.read()
yaml_reader = yaml.load(data)
label_list = yaml_reader['label_list']
print(label_list)
# set parameters of model
model_file = "inference_model/model/model.pdmodel"
params_file = "inference_model/model/model.pdiparams"
# init pur model
predictor = predict_config(model_file, params_file)
cap = cv2.VideoCapture(0)
ret,img = cap.read()
im_size = 320
scale_factor = np.array([im_size* 1./ img.shape[0], im_size * 1./img.shape[1]]).reshape((1,2)).astype(np.float32)
im_shape = np.array([im_size,im_size]).reshape((1,2)).astype(np.float32)
while True:
ret, frame = cap.read()
data = preprocess(frame, im_size)
time_start = time.time()
result = predict(predictor, [im_shape, data, scale_factor])
print('Time Cost: {}'.format(time.time() - time_start))
fps = time.time()-time_start
fps = 1 / fps
draw_bbox_image(frame, result[0], label_list,fps=fps, threshold=0.1)
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
main()
在上面的目录下新建Terminal，运行
python3 infer.py
学大模型，用大模型上飞桨星河社区！每天8点V100G算力免费领！免费领取ERNIE 4.0 100w Token >>>
更多推荐
0
0- 0
已为社区贡献1438条内容
所有评论(0)