caffe 利用VGG训练自己的数据
写这个是因为有童鞋在跑VGG的时候遇到各种问题，供参考一下。
网络结构
以VGG16为例，自己跑的细胞数据
solver.prototxt：
net: "/media/dl/source/Experiment/cell/test/vgg/vgg16.prototxt"
test_iter:
test_interval:
base_lr: 0.0001
lr_policy: "step"
gamma: 0.1
stepsize:
display:
max_iter:
momentum: 0.9
weight_decay: 0.0005
snapshot:
snapshot_prefix: "/media/dl/source/Experiment/cell/test/vgg/vgg"
solver_mode: GPU
vgg16.prototxt：
注意，这里的数据层我是用的“ImageData”格式，也就是没有转为LMDB，直接导入图片进去的，因为我用的服务器，为了方便。如果为了更高效，还是使用LMDB数据库的形式。使用LMDB数据库形式的数据层我也写了下，放在这个prototxt后面作为补充。
另外，注意修改最后一个全连接层的num_output为自己的类别数。并修改该层的名字，如我改为了“cellfc8”，是为了finetune vgg时重新训练该层，不使用该层的预训练参数。
name: "VGG16"
layer {
name: "data"
type: "ImageData"
top: "data"
top: "label"
include {
phase: TRAIN
}
# transform_param {
# mirror: true
# crop_size:
# mean_file: "data/ilsvrc12_shrt_256/imagenet_mean.binaryproto"
# } image_data_param {
source: "/media/dl/source/Experiment/cell/data/trainnew2_resize/trainnew.txt"
batch_size:
shuffle:true
#is_color: false
new_height:
new_width:
}
}
layer {
name: "data"
type: "ImageData"
top: "data"
top: "label"
include {
phase: TEST
}
# transform_param {
# mirror: false
# crop_size:
# mean_file: "data/ilsvrc12_shrt_256/imagenet_mean.binaryproto"
# } image_data_param {
source: "/media/dl/source/Experiment/cell/data/val2_resize/valnew.txt"
batch_size:
#is_color: false
new_height:
new_width:
}
}
layer {
bottom: "data"
top: "conv1_1"
name: "conv1_1"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv1_1"
top: "conv1_1"
name: "relu1_1"
type: "ReLU"
}
layer {
bottom: "conv1_1"
top: "conv1_2"
name: "conv1_2"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv1_2"
top: "conv1_2"
name: "relu1_2"
type: "ReLU"
}
layer {
bottom: "conv1_2"
top: "pool1"
name: "pool1"
type: "Pooling"
pooling_param {
pool: MAX
kernel_size:
stride:
}
}
layer {
bottom: "pool1"
top: "conv2_1"
name: "conv2_1"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv2_1"
top: "conv2_1"
name: "relu2_1"
type: "ReLU"
}
layer {
bottom: "conv2_1"
top: "conv2_2"
name: "conv2_2"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv2_2"
top: "conv2_2"
name: "relu2_2"
type: "ReLU"
}
layer {
bottom: "conv2_2"
top: "pool2"
name: "pool2"
type: "Pooling"
pooling_param {
pool: MAX
kernel_size:
stride:
}
}
layer {
bottom: "pool2"
top: "conv3_1"
name: "conv3_1"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv3_1"
top: "conv3_1"
name: "relu3_1"
type: "ReLU"
}
layer {
bottom: "conv3_1"
top: "conv3_2"
name: "conv3_2"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv3_2"
top: "conv3_2"
name: "relu3_2"
type: "ReLU"
}
layer {
bottom: "conv3_2"
top: "conv3_3"
name: "conv3_3"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv3_3"
top: "conv3_3"
name: "relu3_3"
type: "ReLU"
}
layer {
bottom: "conv3_3"
top: "pool3"
name: "pool3"
type: "Pooling"
pooling_param {
pool: MAX
kernel_size:
stride:
}
}
layer {
bottom: "pool3"
top: "conv4_1"
name: "conv4_1"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv4_1"
top: "conv4_1"
name: "relu4_1"
type: "ReLU"
}
layer {
bottom: "conv4_1"
top: "conv4_2"
name: "conv4_2"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv4_2"
top: "conv4_2"
name: "relu4_2"
type: "ReLU"
}
layer {
bottom: "conv4_2"
top: "conv4_3"
name: "conv4_3"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv4_3"
top: "conv4_3"
name: "relu4_3"
type: "ReLU"
}
layer {
bottom: "conv4_3"
top: "pool4"
name: "pool4"
type: "Pooling"
pooling_param {
pool: MAX
kernel_size:
stride:
}
}
layer {
bottom: "pool4"
top: "conv5_1"
name: "conv5_1"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv5_1"
top: "conv5_1"
name: "relu5_1"
type: "ReLU"
}
layer {
bottom: "conv5_1"
top: "conv5_2"
name: "conv5_2"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv5_2"
top: "conv5_2"
name: "relu5_2"
type: "ReLU"
}
layer {
bottom: "conv5_2"
top: "conv5_3"
name: "conv5_3"
type: "Convolution"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
convolution_param {
num_output:
pad:
kernel_size:
weight_filler {
type: "gaussian"
std: 0.01
}
bias_filler {
type: "constant"
value:
}
}
}
layer {
bottom: "conv5_3"
top: "conv5_3"
name: "relu5_3"
type: "ReLU"
}
layer {
bottom: "conv5_3"
top: "pool5"
name: "pool5"
type: "Pooling"
pooling_param {
pool: MAX
kernel_size:
stride:
}
}
layer {
bottom: "pool5"
top: "fc6"
name: "fc6"
type: "InnerProduct"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
inner_product_param {
num_output:
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
bottom: "fc6"
top: "fc6"
name: "relu6"
type: "ReLU"
}
layer {
bottom: "fc6"
top: "fc6"
name: "drop6"
type: "Dropout"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
bottom: "fc6"
top: "fc7"
name: "fc7"
type: "InnerProduct"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
inner_product_param {
num_output:
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
bottom: "fc7"
top: "fc7"
name: "relu7"
type: "ReLU"
}
layer {
bottom: "fc7"
top: "fc7"
name: "drop7"
type: "Dropout"
dropout_param {
dropout_ratio: 0.5
}
}
layer {
bottom: "fc7"
top: "fc8"
name: "cellfc8"
type: "InnerProduct"
param {
lr_mult:
decay_mult:
}
param {
lr_mult:
decay_mult:
}
inner_product_param {
num_output: 7 #改为自己的类别数
weight_filler {
type: "gaussian"
std: 0.005
}
bias_filler {
type: "constant"
value: 0.1
}
}
}
layer {
name: "accuracy_at_1"
type: "Accuracy"
bottom: "fc8"
bottom: "label"
top: "accuracy_at_1"
accuracy_param {
top_k:
}
include {
phase: TEST
}
}
layer {
name: "accuracy_at_5"
type: "Accuracy"
bottom: "fc8"
bottom: "label"
top: "accuracy_at_5"
accuracy_param {
top_k:
}
include {
phase: TEST
}
}
layer {
bottom: "fc8"
bottom: "label"
top: "loss"
name: "loss"
type: "SoftmaxWithLoss"
}
如果使用LMDB数据库形式，将前面的数据层改为：
name: "vgg"
layer {
name: "data"
type: "Data"
top: "data"
top: "label"
include {
phase: TRAIN
}
transform_param {
mirror: true
crop_size:
#如果图片大于224，则使用crop的方式，小于则使用下面的new_height和new_width
# new_height:
#new_width:
mean_file: "vggface/face_mean.binaryproto"
}
data_param {
source: "vggface/face_train_lmdb"
batch_size:
backend: LMDB
}
}
layer {
name: "data"
type: "Data"
top: "data"
top: "label"
include {
phase: TEST
}
transform_param {
mirror: false
crop_size:
#如果图片大于224，则使用crop的方式，小于则使用下面的new_height和new_width
# new_height:
#new_width:
mean_file: "vggface/face_mean.binaryproto"
}
data_param {
source: "vggface/face_val_lmdb"
batch_size:
backend: LMDB
}
}
训练
放一个shell命令：
#!/usr/bin/env sh TOOLS=/home/dl/caffe-jonlong/build/tools $TOOLS/caffe train \
-solver=/media/dl/source/Experiment/cell/test/vgg/solver.prototxt \
-weights=/media/dl/source/Experiment/cell/test/vgg/VGG_ILSVRC_16_layers.caffemodel \
-gpu=all \
预训练模型VGG_ILSVRC_16_layers.caffemodel的下载地址为
caffe 利用VGG训练自己的数据的更多相关文章
- 利用YOLOV3训练自己的数据
写在前面:YOLOV3只有修改了源码才需要重新make,而且make之前要先make clean. 一.准备数据 在/darknet/VOCdevkit1下建立文件夹VOC2007. voc2007文 ...
- caffe 如何训练自己的数据图片
申明:此教程加工于caffe 如何训练自己的数据图片 一.准备数据 有条件的同学,可以去imagenet的官网http://www.image-net.org/download-images,下载im ...
- caffe学习三：使用Faster RCNN训练自己的数据
本文假设你已经完成了安装,并可以运行demo.py 不会安装且用PASCAL VOC数据集的请看另来两篇博客. caffe学习一:ubuntu16.04下跑Faster R-CNN demo (基于c ...
- caffe 用faster rcnn 训练自己的数据 遇到的问题
1 . 怎么处理那些pyx和.c .h文件 在lib下有一些文件为.pyx文件,遇到不能import可以cython 那个文件,然后把lib文件夹重新make一下. 遇到.c 和 .h一样的操作. 2 ...
- YOLO2解读，训练自己的数据及相关转载以供学习
https://pjreddie.com/darknet/yolo/ 具体安装及使用可以参考官方文档https://github.com/pjreddie/darknet https://blog.c ...
- YOLOv3：训练自己的数据（附优化与问题总结）
环境说明 系统:ubuntu16.04 显卡:Tesla k80 12G显存 python环境: 2.7 && 3.6 前提条件:cuda9.0 cudnn7.0 opencv3.4. ...
- 人脸检测及识别python实现系列（3）——为模型训练准备人脸数据
人脸检测及识别python实现系列(3)——为模型训练准备人脸数据 机器学习最本质的地方就是基于海量数据统计的学习,说白了,机器学习其实就是在模拟人类儿童的学习行为.举一个简单的例子,成年人并没有主动 ...
- TensorFlow下利用MNIST训练模型识别手写数字
本文将参考TensorFlow中文社区官方文档使用mnist数据集训练一个多层卷积神经网络(LeNet5网络),并利用所训练的模型识别自己手写数字. 训练MNIST数据集,并保存训练模型 # Pyth ...
- py-faster-rcnn 训练自己的数据
转载:http://blog.csdn.net/sinat_30071459/article/details/51332084 Faster-RCNN+ZF用自己的数据集训练模型(Python版本) ...
随机推荐
- Web前端的缓存机制（那些以代价换来的效率）
对于Web前端而言,cache可以说是无处不在,通常是2个环节之间,就会引入一个cache做为提升整体效率的角色.例如A和B两者之间的数据交换,为了提升整体的效率,引入角色C,而C被用于当做热点数据的 ...
- SpringBoot启动方式讲解和部署war项目到tomcat9
1.SpringBoot启动方式讲解和部署war项目到tomcat9简介:SpringBoot常见启动方式讲解和部署war项目Tomcat 1.ide启动 2.jar包方式启动 maven插件: &l ...
- 2017-2018-2 20155303『网络对抗技术』Exp8：Web基础
2017-2018-2 『网络对抗技术』Exp8:Web基础 --------CONTENTS-------- 一.原理与实践说明 1.实践具体要求 2.基础问题回答 二.实践过程记录 1.Web前端 ...
- jquery菜单插件
原理很简单. 涉及到知识点: 1.jquery的position注意这里是jquery的position,不是css的position offset的概念 2.>的概念. 3..ulh>l ...
- 如何用ModelsimSE仿真IP核-以PLL为例
我们之前介绍了如何使用Modelsim SE进行仿真和利用do文件的仿真方法,但是其中待仿真的模块是我们自己编写的Verilog模块,但是在实际工作中,我们的设计中会经常用到FPGA厂商给我们提供的现 ...
- 【ARTS】01_03_左耳听风-20181126~1202
ARTS: Algrothm: leetcode算法题目 Review: 阅读并且点评一篇英文技术文章 Tip/Techni: 学习一个技术技巧 Share: 分享一篇有观点和思考的技术文章 Algo ...
- python 历险记（二）— python 的面向对象
目录 前言 类和对象 如何定义和实例化类? 如何定义和使用属性? 什么是方法? 静态方法和普通方法 构造函数该怎么写? str 方法怎么写? 多态是什么? 继承性和 java 是一样的吗? 父类和子类 ...
- 使用 Linux 系统调用的内核命令【转】
转自:http://www.ibm.com/developerworks/cn/linux/l-system-calls/ 探究 SCI 并添加自己的调用 Linux® 系统调用 —— 我们每天都在使 ...
- linux /proc目录说明（访问内核数据结构，修改内核设置）
1. /proc目录 Linux 内核提供了一种通过 /proc 文件系统,在运行时访问内核内部数据结构.改变内核设置的机制.proc文件系统是一个伪文件系统,它只存在内存当中,而不占用外存空间.它以 ...
- servlet请求中的信息
在servlet中HttpServeltRequest中有一个方法getRequestURL() 假如我们平常输入的地址是:localhost/Demo1/TestServlet?hello=worl ...