当前位置： 首页 > news >正文

PaddleVideo：Squeeze Time算法移植

news 来源：原创 2024/9/19 10:05:21

 参考
PaddleVideo/docs/zh-CN/contribute/add_new_algorithm.md at develop · PaddlePaddle/PaddleVideo · GitHubAwesome video understanding toolkits based on PaddlePaddle. It supports video data annotation tools, lightweight RGB and skeleton based action recognition model, practical applications for video tagging and sport action detection. - PaddleVideo/docs/zh-CN/contribute/add_new_algorithm.md at develop · PaddlePaddle/PaddleVideoicon-default.png?t=N7T8https://github.com/PaddlePaddle/PaddleVideo/blob/develop/docs/zh-CN/contribute/add_new_algorithm.md

1：添加backbone：（网络我自己砍了几刀，目的是想和ppTSM-v2做对比）

paddlevideo/modeling/backbones/squeezetime.py

from __future__ import absolute_import, division, print_functionimport paddle
import paddle.nn as nn
from paddle import ParamAttr
from paddle.nn import AdaptiveAvgPool2D, BatchNorm, Conv2D, Dropout, Linear, BatchNorm2D
from paddle.regularizer import L2Decay
from paddle.nn.initializer import KaimingNormal,Constant
import paddle.nn.functional as Ffrom ..registry import BACKBONESdef get_inplanes():return [64, 128, 256, 512]class SpatialConv(nn.Layer):"""Inter-temporal Object Interaction Module (IOI)"""def __init__(self, dim_in, dim_out, pos_dim=7):super(SpatialConv, self).__init__()self.short_conv = nn.Conv2D(dim_in, dim_out, kernel_size=3, stride=1, padding=1, groups=1)self.glo_conv = nn.Sequential(nn.Conv2D(dim_in, 16, kernel_size=3, stride=1, padding=1, groups=1),nn.BatchNorm2D(16), nn.ReLU(),nn.Conv2D(16, 16, kernel_size=7, stride=1, padding=3),nn.BatchNorm2D(16), nn.ReLU(),nn.Conv2D(16, dim_out, kernel_size=3, stride=1, padding=1, groups=1), nn.Sigmoid())self.pos_embed = self.create_parameter(shape=[1, 16, pos_dim, pos_dim], default_initializer=nn.initializer.KaimingNormal())def forward(self, x, param):x_short = self.short_conv(x)x = x * paramfor i in range(len(self.glo_conv)):if i == 3:_, _, H, W = x.shapeif self.pos_embed.shape[2] != H or self.pos_embed.shape[3] != W:pos_embed = F.interpolate(self.pos_embed, size=(H, W), mode='bilinear', align_corners=True)else:pos_embed = self.pos_embedx = x + pos_embedx = self.glo_conv[i](x)return x_short * xclass Conv2d(nn.Layer):"""Channel-Time Learning Module (CTL)"""def __init__(self,in_channels: int,out_channels: int,kernel_size: int,stride: int = 1,padding: int = 0,dilation: int = 1,groups: int = 1,bias: bool = True,padding_mode: str = 'zeros', pos_dim = 7):super(Conv2d, self).__init__()self.stride = strideself.param_conv = nn.Sequential(nn.AdaptiveAvgPool2D((1, 1)),nn.Conv2D(in_channels, in_channels, 1, stride=1, padding=1 // 2, bias_attr=False),nn.BatchNorm2D(in_channels),nn.ReLU(),nn.Conv2D(in_channels, in_channels, 1, bias_attr=False),nn.Sigmoid())self.temporal_conv = nn.Conv2D(in_channels=in_channels, out_channels=out_channels, kernel_size=kernel_size, stride=1, padding=padding, dilation=dilation, groups=groups, bias_attr=bias, padding_mode=padding_mode)self.spatial_conv = SpatialConv(dim_in=in_channels, dim_out=out_channels, pos_dim=pos_dim)def forward(self, x):param = self.param_conv(x)x = self.temporal_conv(param * x) + self.spatial_conv(x, param)return xdef conv3x3x3(in_planes, out_planes, stride=1, pos_dim=7):return Conv2d(in_planes, out_planes, kernel_size=1, stride=stride, padding=0, bias=False, pos_dim=pos_dim)def conv1x1x1(in_planes, out_planes, stride=1):return nn.Conv2D(in_planes, out_planes, kernel_size=1, stride=stride, bias_attr=False)class BasicBlock(nn.Layer):"""Channel-Time Learning (CTL) Block"""expansion = 1def __init__(self, in_planes, planes, stride=1, shortcut_conv=None, pos_dim=7):super().__init__()self.conv1 = conv3x3x3(in_planes, planes, stride)self.bn1 = nn.BatchNorm2D(planes)self.relu = nn.ReLU()self.conv2 = conv3x3x3(planes, planes, pos_dim=pos_dim)self.bn2 = nn.BatchNorm2D(planes)self.shortcut_conv = shortcut_convself.stride = strideif stride != 1:self.downsample = nn.Sequential(nn.Conv2D(in_planes, in_planes, kernel_size=2, stride=2, groups=in_planes),nn.BatchNorm2D(in_planes))def forward(self, x):if self.stride != 1:x = self.downsample(x)residual = xout = self.conv1(x)out = self.bn1(out)out = self.relu(out)out = self.conv2(out)out = self.bn2(out)if self.shortcut_conv is not None:residual = self.shortcut_conv(x)out += residualout = self.relu(out)return outclass Bottleneck(nn.Layer):"""Channel-Time Learning (CTL) Block"""expansion = 4def __init__(self, in_planes, planes, stride=1, shortcut_conv=None, pos_dim=7):super().__init__()self.conv1 = conv1x1x1(in_planes, planes)self.bn1 = nn.BatchNorm2D(planes)self.conv2 = conv3x3x3(planes, planes, pos_dim=pos_dim)self.bn2 = nn.BatchNorm2D(planes)self.conv3 = conv1x1x1(planes, planes * self.expansion)self.bn3 = nn.BatchNorm2D(planes * self.expansion)self.relu = nn.ReLU()self.shortcut_conv = shortcut_convself.stride = strideif stride != 1:self.downsample = nn.Sequential(nn.Conv2D(in_planes, in_planes, kernel_size=2, stride=2, groups=in_planes),nn.BatchNorm2D(in_planes))def forward(self, x):if self.stride != 1:x = self.downsample(x)residual = xout = self.conv1(x)out = self.bn1(out)out = self.relu(out)out = self.conv2(out)out = self.bn2(out)out = self.relu(out)out = self.conv3(out)out = self.bn3(out)if self.shortcut_conv is not None:residual = self.shortcut_conv(x)out += residualout = self.relu(out)return outclass ResNet(nn.Layer):def __init__(self,block,layers,block_inplanes,n_input_channels=3,no_max_pool=False,shortcut_type='B',widen_factor=1.0,dropout=0.2, freeze_bn=False, spatial_stride=[1,2,2,2], pos_dim=[64,32,16,8]):super().__init__()self.freeze_bn = freeze_bnblock_inplanes = [int(x * widen_factor) for x in block_inplanes]self.in_planes = block_inplanes[0]self.no_max_pool = no_max_poolself.dropout = dropoutself.conv1 = nn.Conv2D(n_input_channels,self.in_planes,kernel_size=5,stride=2,padding=2,groups=1,bias_attr=False)self.bn1 = nn.BatchNorm2D(self.in_planes)self.relu = nn.ReLU()self.maxpool = nn.MaxPool2D(kernel_size=3, stride=2, padding=1)self.layer1 = self._make_layer(block, block_inplanes[0], layers[0],shortcut_type, stride=spatial_stride[0], pos_dim=pos_dim[0])self.layer2 = self._make_layer(block,block_inplanes[1],layers[1],shortcut_type,stride=spatial_stride[1], pos_dim=pos_dim[1])self.layer3 = self._make_layer(block,block_inplanes[2],layers[2],shortcut_type,stride=spatial_stride[2], pos_dim=pos_dim[2])self.layer4 = self._make_layer(block,block_inplanes[3],layers[3],shortcut_type,stride=spatial_stride[3], pos_dim=pos_dim[3])def _downsample_basic_block(self, x, planes, stride):out = F.avg_pool2d(x, kernel_size=1, stride=stride)zero_pads = paddle.zeros([out.shape[0], planes - out.shape[1], out.shape[2], out.shape[3]])if isinstance(out, paddle.CUDAPlace):zero_pads = zero_pads.cuda()out = paddle.concat([out, zero_pads], axis=1)return outdef _make_layer(self, block, planes, blocks, shortcut_type, stride=1, pos_dim=7):shortcut = Noneif self.in_planes != planes * block.expansion:shortcut = nn.Sequential(conv1x1x1(self.in_planes, planes * block.expansion, stride=1),nn.BatchNorm2D(planes * block.expansion))layers = []layers.append(block(in_planes=self.in_planes,planes=planes,stride=stride, shortcut_conv=shortcut, pos_dim=pos_dim))self.in_planes = planes * block.expansionfor i in range(1, blocks):layers.append(block(self.in_planes, planes, pos_dim=pos_dim))return nn.Sequential(*layers)def forward(self, x):print('##################', x.shape)if len(x.shape) == 3:x = paddle.unsqueeze(x, axis=0)N, C, H, W = x.shapex = x.reshape([int(N/16), -1, H, W])x = self.conv1(x)x = self.bn1(x)x = self.relu(x)if not self.no_max_pool:x = self.maxpool(x)x = self.layer1(x)x = self.layer2(x)x = self.layer3(x)x = self.layer4(x)return xdef train(self, mode=True):freeze_bn = self.freeze_bnfreeze_bn_affine = self.freeze_bnsuper(ResNet, self).train(mode)if freeze_bn:print("Freezing Mean/Var of BatchNorm2D.")for m in self.sublayers():if isinstance(m, nn.BatchNorm2D):m.eval()if freeze_bn_affine:print("Freezing Weight/Bias of BatchNorm2D.")for m in self.sublayers():if isinstance(m, nn.BatchNorm2D):m.weight.stop_gradient = Truem.bias.stop_gradient = Truedef SqueezeTime_model(**kwargs):model = ResNet(Bottleneck, [2, 2, 2, 2], get_inplanes(), **kwargs)return model@BACKBONES.register()
def SqueezeTime(pretrained=None, use_ssld=False, **kwargs):"""Build SqueezeTime Model"""model = SqueezeTime_model(widen_factor=0.5, dropout=0.5, n_input_channels=48, freeze_bn=False, spatial_stride=[1, 2, 2, 2], pos_dim=[64, 32, 16, 8])return  model

2：导入backbone：

paddlevideo/modeling/backbones/__init__.py

# Copyright (c) 2020  PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.from .actbert import BertForMultiModalPreTraining
from .adds import ADDS_DepthNet
from .agcn import AGCN
from .asrf import ASRF
from .bmn import BMN
from .cfbi import CFBI
from .movinet import MoViNet
from .ms_tcn import MSTCN
from .resnet import ResNet
from .resnet_slowfast import ResNetSlowFast
from .resnet_slowfast_MRI import ResNetSlowFast_MRI
from .resnet_tsm import ResNetTSM
from .resnet_tsm_MRI import ResNetTSM_MRI
from .resnet_tsn_MRI import ResNetTSN_MRI
from .resnet_tweaks_tsm import ResNetTweaksTSM
from .resnet_tweaks_tsn import ResNetTweaksTSN
from .stgcn import STGCN
from .swin_transformer import SwinTransformer3D
from .transnetv2 import TransNetV2
from .vit import VisionTransformer
from .vit_tweaks import VisionTransformer_tweaks
from .ms_tcn import MSTCN
from .asrf import ASRF
from .resnet_tsn_MRI import ResNetTSN_MRI
from .resnet_tsm_MRI import ResNetTSM_MRI
from .resnet_slowfast_MRI import ResNetSlowFast_MRI
from .cfbi import CFBI
from .ctrgcn import CTRGCN
from .agcn2s import AGCN2s
from .movinet import MoViNet
from .resnet3d_slowonly import ResNet3dSlowOnly
from .toshift_vit import TokenShiftVisionTransformer
from .pptsm_mv2 import PPTSM_MobileNetV2
from .pptsm_mv3 import PPTSM_MobileNetV3
from .pptsm_v2 import PPTSM_v2
from .yowo import YOWO
from .squeezetime import SqueezeTime__all__ = ['ResNet', 'ResNetTSM', 'ResNetTweaksTSM', 'ResNetSlowFast', 'BMN','ResNetTweaksTSN', 'VisionTransformer', 'STGCN', 'AGCN', 'TransNetV2','ADDS_DepthNet', 'VisionTransformer_tweaks', 'BertForMultiModalPreTraining','ResNetTSN_MRI', 'ResNetTSM_MRI', 'ResNetSlowFast_MRI', 'CFBI', 'MSTCN','ASRF', 'MoViNet', 'SwinTransformer3D', 'CTRGCN','TokenShiftVisionTransformer', 'AGCN2s', 'PPTSM_MobileNetV2','PPTSM_MobileNetV3', 'PPTSM_v2', 'ResNet3dSlowOnly', 'YOWO', 'SqueezeTime'
]

3：添加head：

paddlevideo/modeling/heads/i2d_head.py

# Copyright (c) 2020  PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.import paddle
import paddle.nn as nn
from paddle import ParamAttrfrom ..registry import HEADS
from ..weight_init import weight_init_
from .base import BaseHead@HEADS.register()
class I2DHead(BaseHead):"""Classification head for I2D.Args:num_classes (int): Number of classes to be classified.in_channels (int): Number of channels in input feature.loss_cls (dict): Config for building loss.Default: dict(name='CrossEntropyLoss')spatial_type (str): Pooling type in spatial dimension. Default: 'avg'.drop_ratio (float): Probability of dropout layer. Default: 0.5.std (float): Std value for Initiation. Default: 0.01.kwargs (dict, optional): Any keyword argument to be used to initializethe head."""def __init__(self,num_classes,in_channels,loss_cfg=dict(name='CrossEntropyLoss'),spatial_type='avg',drop_ratio=0.5,std=0.01,**kwargs):super().__init__(num_classes, in_channels, loss_cfg, **kwargs)self.spatial_type = spatial_typeself.dropout_ratio = drop_ratioself.init_std = stdif self.dropout_ratio != 0:self.dropout = nn.Dropout(p=self.dropout_ratio)else:self.dropout = Noneself.fc_cls = nn.Linear(self.in_channels, self.num_classes)if self.spatial_type == 'avg':self.avg_pool = nn.AdaptiveAvgPool2D((1, 1))else:self.avg_pool = nn.AdaptiveMaxPool2D((1,1))def forward(self, x, num_segs = None):"""Defines the computation performed at every call.Args:x (Tensor): The input data.Returns:Tensor: The classification scores for input samples."""   # [N, in_channels, 4, 7, 7]if self.avg_pool is not None:x = self.avg_pool(x)# [N, in_channels, 1, 1, 1]if self.dropout is not None:x = self.dropout(x)# [N, in_channels, 1, 1, 1]x = paddle.reshape(x, [x.shape[0], -1])# [N, in_channels]cls_score = self.fc_cls(x)# [N, num_classes]return cls_score# def forward_new(self, x, num_segs = None):#     """Defines the computation performed at every call.#     Args:#         x (Tensor): The input data.#     Returns:#         Tensor: The classification scores for input samples.#     """      #     # [N, in_channels, 4, 7, 7]#     if self.avg_pool is not None:#         x = self.avg_pool(x)#     # [N, in_channels, 1, 1, 1]#     if self.dropout is not None:#         x = self.dropout(x)#     # [N, in_channels, 1, 1, 1]#     x = paddle.reshape(x, [x.shape[0], -1])#     # [N, in_channels]#     cls_score = self.fc_cls(x)#     # [N, num_classes]#     return cls_score

4：导入head：

paddlevideo/modeling/heads/__init__.py

# Copyright (c) 2020  PaddlePaddle Authors. All Rights Reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License"
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.from .adds_head import AddsHead
from .asrf_head import ASRFHead
from .attention_lstm_head import AttentionLstmHead, ActionAttentionLstmHead
from .base import BaseHead
from .bbox_head import BBoxHeadAVA
from .cfbi_head import CollaborativeEnsemblerMS
from .i3d_head import I3DHead
from .movinet_head import MoViNetHead
from .ms_tcn_head import MSTCNHead
from .pptimesformer_head import ppTimeSformerHead
from .pptsm_head import ppTSMHead
from .pptsn_head import ppTSNHead
from .roi_head import AVARoIHead
from .single_straight3d import SingleRoIExtractor3D
from .slowfast_head import SlowFastHead
from .stgcn_head import STGCNHead
from .timesformer_head import TimeSformerHead
from .transnetv2_head import TransNetV2Head
from .tsm_head import TSMHead
from .tsn_head import TSNHead
from .ms_tcn_head import MSTCNHead
from .asrf_head import ASRFHead
from .ctrgcn_head import CTRGCNHead
from .movinet_head import MoViNetHead
from .agcn2s_head import AGCN2sHead
from .token_shift_head import TokenShiftHead
from .i2d_head import I2DHead__all__ = ['BaseHead', 'TSNHead', 'TSMHead', 'ppTSMHead', 'ppTSNHead', 'SlowFastHead','AttentionLstmHead', 'TimeSformerHead', 'STGCNHead', 'TransNetV2Head','I3DHead', 'SingleRoIExtractor3D', 'AVARoIHead', 'BBoxHeadAVA', 'AddsHead','ppTimeSformerHead', 'CollaborativeEnsemblerMS', 'MSTCNHead', 'ASRFHead','MoViNetHead', 'CTRGCNHead', 'TokenShiftHead', 'ActionAttentionLstmHead','AGCN2sHead', 'I2DHead'
]

5：训练配置文件：

configs/recognition/pptsm/v2/md_ppsqt_16frames_uniform.yaml

MODEL: #MODEL fieldframework: "Recognizer2D" #Mandatory, indicate the type of network, associate to the 'paddlevideo/modeling/framework/' .backbone: #Mandatory, indicate the type of backbone, associate to the 'paddlevideo/modeling/backbones/' .name: "SqueezeTime" #Mandatory, The name of backbone.head:name: "I2DHead" #Mandatory, indicate the type of head, associate to the 'paddlevideo/modeling/heads'#pretrained: "" #Optional, pretrained model path.num_classes: 2in_channels: 1024DATASET: #DATASET fieldbatch_size: 16  #Mandatory, bacth sizenum_workers: 4 #Mandatory, the number of subprocess on each GPU.train:format: "FrameDataset" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset'data_prefix: "/home/mnt/sdd/Data/data_fights/rawframes" #Mandatory, train data root pathfile_path: "/home/mnt/sdd/Data/data_fights/train_list.txt" #Mandatory, train data index file pathsuffix: 'img_{:06}.jpg'valid:format: "FrameDataset" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset'data_prefix: "/home/mnt/sdd/Data/data_fights/rawframes" #Mandatory, valid data root pathfile_path: "/home/mnt/sdd/Data/data_fights/test_list.txt" #Mandatory, valid data index file pathsuffix: 'img_{:06}.jpg'test:format: "FrameDataset" #Mandatory, indicate the type of dataset, associate to the 'paddlevidel/loader/dateset'data_prefix: "/home/mnt/sdd/Data/data_fights/rawframes" #Mandatory, valid data root pathfile_path: "/home/mnt/sdd/Data/data_fights/test_list.txt" #Mandatory, valid data index file pathsuffix: 'img_{:06}.jpg'PIPELINE: #PIPELINE fieldtrain: #Mandotary, indicate the pipeline to deal with the training data, associate to the 'paddlevideo/loader/pipelines/'decode:name: "FrameDecoder"sample:name: "Sampler"num_seg: 16seg_len: 1valid_mode: Falsetransform: #Mandotary, image transfrom operator- Scale:short_size: 256- MultiScaleCrop:target_size: 256- RandomCrop:target_size: 224- RandomFlip:- Image2Array:- Normalization:mean: [0.485, 0.456, 0.406]std: [0.229, 0.224, 0.225]valid: #Mandatory, indicate the pipeline to deal with the validing data. associate to the 'paddlevideo/loader/pipelines/'decode:name: "FrameDecoder"sample:name: "Sampler"num_seg: 16seg_len: 1valid_mode: Truetransform:- Scale:short_size: 256- CenterCrop:target_size: 224- Image2Array:- Normalization:mean: [0.485, 0.456, 0.406]std: [0.229, 0.224, 0.225]test:  #Mandatory, indicate the pipeline to deal with the validing data. associate to the 'paddlevideo/loader/pipelines/'decode:name: "FrameDecoder"sample:name: "Sampler"num_seg: 16seg_len: 1valid_mode: Truetransform:- Scale:short_size: 256- CenterCrop:target_size: 224- Image2Array:- Normalization:mean: [0.485, 0.456, 0.406]std: [0.229, 0.224, 0.225]OPTIMIZER: #OPTIMIZER fieldname: 'Momentum'momentum: 0.9learning_rate:iter_step: Truename: 'CustomWarmupCosineDecay'max_epoch: 120warmup_epochs: 10warmup_start_lr: 0.005cosine_base_lr: 0.01weight_decay:name: 'L2'value: 1e-4use_nesterov: TrueMIX:name: "Mixup"alpha: 0.2METRIC:name: 'CenterCropMetric'INFERENCE:name: 'ppSQT_Inference_helper'num_seg: 16target_size: 224model_name: "ppSQT"
log_interval: 10 #Optional, the interal of logger, default:10
epochs: 120  #Mandatory, total epoch
log_level: "INFO" #Optional, the logger level. default: "INFO"

6：训练：

# multi-gpu-st
export CUDA_VISIBLE_DEVICES=0,1
python -B -m paddle.distributed.launch --gpus="0,1"  --log_dir=./log/log_sqt_frame_16  main.py  --validate -c configs/recognition/pptsm/v2/ppsqt_lcnet_md_16frames_uniform.yaml

7：结果：精度比ppTSM-v2低8个点左右。有可能是没有预训练权重的问题。 

相关文章：

  • 北京网站建设多少钱？
  • 辽宁网页制作哪家好_网站建设
  • 高端品牌网站建设_汉中网站制作
  • WebOffice在线编微软Offfice，并以二进制流的形式打开Word文档
  • 【软件测试】 1+X初级 功能测试试题
  • antDesignPro随记
  • Python OpenCV 教学取得视频资讯
  • 电动卡丁车语音芯片方案选型：让驾驶体验更智能、更安全
  • el-date-picker 禁用 之前 和 之后 的时间
  • uniapp easycom组件冲突
  • 云服务部署和管理：容器化与微服务的融合之道
  • mybatisplus不想输出批量插入和查询日志
  • 观察矩阵(View Matrix)、投影矩阵(Projection Matrix)、视口矩阵(Window Matrix)及VPM矩阵及它们之间的关系
  • 微服务-初级篇
  • mac｜Mac压缩与解压缩
  • 在程序中引用cuda.memory函数监控GPU内存
  • Linux系统编程——线程控制
  • 聚道云赋能，财务流程大提速，NC凭证对接一键通！
  • [PHP内核探索]PHP中的哈希表
  • ［分享］iOS开发－关于在xcode中引用文件夹右边出现问号的解决办法
  • 《微软的软件测试之道》成书始末、出版宣告、补充致谢名单及相关信息
  • Android框架之Volley
  • CSS中外联样式表代表的含义
  • express + mock 让前后台并行开发
  • Git同步原始仓库到Fork仓库中
  • JavaScript/HTML5图表开发工具JavaScript Charts v3.19.6发布【附下载】
  • javascript面向对象之创建对象
  • Linux CTF 逆向入门
  • python学习笔记 - ThreadLocal
  • Spark VS Hadoop：两大大数据分析系统深度解读
  • SpringBoot 实战 (三) | 配置文件详解
  • 分享一个自己写的基于canvas的原生js图片爆炸插件
  • 湖南卫视：中国白领因网络偷菜成当代最寂寞的人?
  • 基于Mobx的多页面小程序的全局共享状态管理实践
  • 前端路由实现-history
  • 十年未变！安全，谁之责？（下）
  • 使用Envoy 作Sidecar Proxy的微服务模式-4.Prometheus的指标收集
  • 新手搭建网站的主要流程
  • ![CDATA[ ]] 是什么东东
  • （￥1011）－（一千零一拾一元整）输出
  • (8)Linux使用C语言读取proc/stat等cpu使用数据
  • (C#)获取字符编码的类
  • (ResultSet.TYPE_SCROLL_INSENSITIVE,ResultSet.CONCUR_READ_ONLY)讲解
  • （第27天）Oracle 数据泵转换分区表
  • (附源码)springboot课程在线考试系统 毕业设计 655127
  • （十二）devops持续集成开发——jenkins的全局工具配置之sonar qube环境安装及配置
  • (转贴)用VML开发工作流设计器 UCML.NET工作流管理系统
  • ****** 二十三 ******、软设笔记【数据库】-数据操作-常用关系操作、关系运算
  • .equal()和==的区别 怎样判断字符串为空问题： Illegal invoke-super to void nio.file.AccessDeniedException
  • .helper勒索病毒的最新威胁：如何恢复您的数据？
  • .Mobi域名介绍
  • .mysql secret在哪_MySQL如何使用索引
  • .NET 4 并行（多核）“.NET研究”编程系列之二 从Task开始
  • .Net Core webapi RestFul 统一接口数据返回格式
  • .net MVC中使用angularJs刷新页面数据列表
  • .Net OpenCVSharp生成灰度图和二值图
  • .NET 命令行参数包含应用程序路径吗？
  • .net6解除文件上传限制。Multipart body length limit 16384 exceeded