使用caffe (create_imagenet.sh)将我的jpg转换为lmdb时的问题

Question

我想做一些微调来训练我的jpg。在大量文档的指导下，我主要修改了create_imagenet.sh中的data、txt和tool的路径，具体如下：

#!/usr/bin/env sh
# This script converts the mnist data into lmdb/leveldb format,
# depending on the value assigned to $BACKEND.
set -e

EXAMPLE=/home/sun/Documents/python/fine-tuning/Oxford102/data/test
DATA_TEST=/home/sun/Documents/python/fine-tuning/Oxford102/data/test/test/
DATA_TRAIN=/home/sun/Documents/python/fine-tuning/Oxford102/data/test/train/
NOTI=/home/sun/Documents/python/fine-tuning/Oxford102/data/test
BUILD=/home/sun/Documents/caffe/build/tools

#BACKEND="lmdb"

# Set RESIZE=true to resize the images to 256x256. Leave as false if images have  
# already been resized using another tool.  
RESIZE=true
if $RESIZE; then  
  RESIZE_HEIGHT=256  
  RESIZE_WIDTH=256  
else  
  RESIZE_HEIGHT=0  
  RESIZE_WIDTH=0  
fi  


if [ ! -d "$DATA_TRAIN" ]; then  
  echo "Error: TRAIN_DATA_ROOT is not a path to a directory: $DATA_TRAIN"  
  echo "Set the TRAIN_DATA_ROOT variable in create_imagenet.sh to the path" \  
       "where the ImageNet training data is stored."  
  exit 1  
fi  

if [ ! -d "$DATA_TEST" ]; then  
  echo "Error: TRAIN_DATA_ROOT is not a path to a directory: $DATA_TEST"  
  echo "Set the TRAIN_DATA_ROOT variable in create_imagenet.sh to the path" \  
       "where the ImageNet training data is stored."  
  exit 1  
fi  

echo "Creating train lmdb..."
GLOG_logtostderr=1 $BUILD/convert_imageset \
    --resize_height=$RESIZE_HEIGHT \ 
    --resize_width=$RESIZE_WIDTH \
    --shuffle \
    $DATA_TRAIN \  
    $NOTI/info_train.txt \  
    $EXAMPLE/img_train_lmdb  

echo "Creating test lmdb..."
GLOG_logtostderr=1 $BUILD/convert_imageset \
    --resize_height=$RESIZE_HEIGHT \ 
    --resize_width=$RESIZE_WIDTH \
    --shuffle \
    $DATA_TEST \  
    $NOTI/info_test.txt \  
    $EXAMPLE/img_test_lmdb

echo "Done."

然后我在终端中运行'sudo sh create_oxford.sh‘，但是什么也得不到。

running result in shell

此外，我的文档结构如下：

在/home/sun/Documents/python/fine-tuning/Oxford102/data/test，中，我放入了info_test.txt、info_train.txt、test(jpg表示测试)和train(jpg表示火车)。在txt中，信息类似于'image_0001.jpg 0‘。

Answer 1

1)我成功地运行了代码。

2)当命令行参数没有正确传递时，我得到了用法但什么也得不到。

3)您可以使用以下命令来调试您的-vx脚本: set shell它将显示如下所示的实际论点：

调整高度-- +/home/u5652/.conda/envs/riCffAlex3.6/bin/convert_imageset _=256--调整大小_宽度=256

convert_imageset:将一组图像转换为leveldb/lmdb格式，用作Caffe的输入。用法: convert_imageset标志ROOTFOLDER/ LISTFILE DB_NAME .....................

根据caffe/tools/convert_imageset.cpp：

if (argc < 4) {
    gflags::ShowUsageWithFlagsRestrict(argv[0], "tools/convert_imageset");
    return 1;
  }

最小参数应该是4，否则它将显示用法。由于参数不正确，您一定一直在获取用法，但什么也得不到。

4)我运行了以下代码，并成功创建了lmdb：

#!/usr/bin/env sh
set -vx
EXAMPLE=examples/imagenet/test
DATA=data/ilsvrc12/test
TOOLS=~/.conda/envs/riCffAlex3.6/bin
TRAIN_DATA_ROOT=~/caffe/data/ilsvrc12/test/train/
VAL_DATA_ROOT=~/caffe/data/ilsvrc12/test/test/

RESIZE=true
if $RESIZE; then
  RESIZE_HEIGHT=256
  RESIZE_WIDTH=256
else
  RESIZE_HEIGHT=0
  RESIZE_WIDTH=0
fi

if [ ! -d "$TRAIN_DATA_ROOT" ]; then
  echo "Error: TRAIN_DATA_ROOT is not a path to a directory: $TRAIN_DATA_ROOT"
  echo "Set the TRAIN_DATA_ROOT variable in create_imagenet.sh to the path" \
       "where the ImageNet training data is stored."
  exit 1
fi

if [ ! -d "$VAL_DATA_ROOT" ]; then
  echo "Error: VAL_DATA_ROOT is not a path to a directory: $VAL_DATA_ROOT"
  echo "Set the VAL_DATA_ROOT variable in create_imagenet.sh to the path" \
       "where the ImageNet validation data is stored."
  exit 1
fi

echo "Creating train lmdb..."

GLOG_logtostderr=1 $TOOLS/convert_imageset \
    --resize_height=$RESIZE_HEIGHT \
    --resize_width=$RESIZE_WIDTH \
    --shuffle \
    $TRAIN_DATA_ROOT \
    $DATA/train.txt \
    $EXAMPLE/train_lmdb

echo "Creating test lmdb..."

GLOG_logtostderr=1 $TOOLS/convert_imageset \
    --resize_height=$RESIZE_HEIGHT \
    --resize_width=$RESIZE_WIDTH \
    --shuffle \
    $VAL_DATA_ROOT \
    $DATA/val.txt \
    $EXAMPLE/test_lmdb

echo "Done."

5)我把细节放在和你放在.txt文件里一样的格式。我输入: TestJpg.jpg 0

注意：图像的名称和扩展名应该与您放入.txt中的名称和扩展名相同

试试这条路。您的错误应该得到解决。

谢谢