switch examples to lmdb (except for custom data loaders)

examples/cifar10/create_cifar10.sh

@@ -13,6 +13,6 @@ rm -rf $EXAMPLE/cifar10_train_leveldb $EXAMPLE/cifar10_test_leveldb
 echo "Computing image mean..."
 
 ./build/tools/compute_image_mean $EXAMPLE/cifar10_train_leveldb \
-  $EXAMPLE/mean.binaryproto
+  $EXAMPLE/mean.binaryproto leveldb
 
 echo "Done."

examples/imagenet/create_imagenet.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env sh
-# Create the imagenet leveldb inputs
+# Create the imagenet lmdb inputs
 # N.B. set the path to the imagenet train + val data dirs
 
 EXAMPLE=examples/imagenet
@@ -34,24 +34,24 @@ if [ ! -d "$VAL_DATA_ROOT" ]; then
   exit 1
 fi
 
-echo "Creating train leveldb..."
+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/ilsvrc12_train_leveldb
+    $EXAMPLE/ilsvrc12_train_lmdb
 
-echo "Creating val leveldb..."
+echo "Creating val lmdb..."
 
 GLOG_logtostderr=1 $TOOLS/convert_imageset \
     --resize_height=$RESIZE_HEIGHT \
     --resize_width=$RESIZE_WIDTH \
     --shuffle \
     $VAL_DATA_ROOT \
     $DATA/val.txt \
-    $EXAMPLE/ilsvrc12_val_leveldb
+    $EXAMPLE/ilsvrc12_val_lmdb
 
 echo "Done."

examples/mnist/convert_mnist_data.cpp

@@ -1,6 +1,5 @@
-//
-// This script converts the MNIST dataset to the leveldb format used
-// by caffe to perform classification.
+// This script converts the MNIST dataset to a lmdb (default) or
+// leveldb (--backend=leveldb) format used by caffe to load data.
 // Usage:
 //    convert_mnist_data [FLAGS] input_image_file input_label_file
 //                        output_db_file
@@ -176,7 +175,7 @@ int main(int argc, char** argv) {
 #endif
 
   gflags::SetUsageMessage("This script converts the MNIST dataset to\n"
-        "the leveldb/lmdb format used by Caffe to perform classification.\n"
+        "the lmdb/leveldb format used by Caffe to load data.\n"
         "Usage:\n"
         "    convert_mnist_data [FLAGS] input_image_file input_label_file "
         "output_db_file\n"

examples/mnist/create_mnist.sh

@@ -1,5 +1,5 @@
 #!/usr/bin/env sh
-# This script converts the mnist data into leveldb/lmdb format,
+# This script converts the mnist data into lmdb/leveldb format,
 # depending on the value assigned to $BACKEND.
 
 EXAMPLE=examples/mnist

examples/mnist/readme.md

@@ -19,7 +19,7 @@ You will first need to download and convert the data format from the MNIST websi
     cd $CAFFE_ROOT/examples/mnist
     ./create_mnist.sh
 
-If it complains that `wget` or `gunzip` are not installed, you need to install them respectively. After running the script there should be two datasets, `mnist-train-leveldb`, and `mnist-test-leveldb`.
+If it complains that `wget` or `gunzip` are not installed, you need to install them respectively. After running the script there should be two datasets, `mnist_train_lmdb`, and `mnist_test_lmdb`.
 
 ## LeNet: the MNIST Classification Model
 
@@ -37,21 +37,22 @@ Specifically, we will write a `caffe::NetParameter` (or in python, `caffe.proto.
 
 ### Writing the Data Layer
 
-Currently, we will read the MNIST data from the leveldb we created earlier in the demo. This is defined by a data layer:
+Currently, we will read the MNIST data from the lmdb we created earlier in the demo. This is defined by a data layer:
 
     layers {
       name: "mnist"
       type: DATA
       data_param {
-        source: "mnist-train-leveldb"
+        source: "mnist_train_lmdb"
+        backend: LMDB
         batch_size: 64
         scale: 0.00390625
       }
       top: "data"
       top: "label"
     }
 
-Specifically, this layer has name `mnist`, type `data`, and it reads the data from the given leveldb source. We will use a batch size of 64, and scale the incoming pixels so that they are in the range \[0,1\). Why 0.00390625? It is 1 divided by 256. And finally, this layer produces two blobs, one is the `data` blob, and one is the `label` blob.
+Specifically, this layer has name `mnist`, type `data`, and it reads the data from the given lmdb source. We will use a batch size of 64, and scale the incoming pixels so that they are in the range \[0,1\). Why 0.00390625? It is 1 divided by 256. And finally, this layer produces two blobs, one is the `data` blob, and one is the `label` blob.
 
 ### Writing the Convolution Layer
 

models/bvlc_alexnet/train_val.prototxt

@@ -5,7 +5,8 @@ layers {
   top: "data"
   top: "label"
   data_param {
-    source: "examples/imagenet/ilsvrc12_train_leveldb"
+    source: "examples/imagenet/ilsvrc12_train_lmdb"
+    backend: LMDB
     batch_size: 256
   }
   transform_param {
@@ -21,7 +22,8 @@ layers {
   top: "data"
   top: "label"
   data_param {
-    source: "examples/imagenet/ilsvrc12_val_leveldb"
+    source: "examples/imagenet/ilsvrc12_val_lmdb"
+    backend: LMDB
     batch_size: 50
   }
   transform_param {

models/bvlc_reference_caffenet/train_val.prototxt

@@ -5,7 +5,8 @@ layers {
   top: "data"
   top: "label"
   data_param {
-    source: "examples/imagenet/ilsvrc12_train_leveldb"
+    source: "examples/imagenet/ilsvrc12_train_lmdb"
+    backend: LMDB
     batch_size: 256
   }
   transform_param {
@@ -21,7 +22,8 @@ layers {
   top: "data"
   top: "label"
   data_param {
-    source: "examples/imagenet/ilsvrc12_val_leveldb"
+    source: "examples/imagenet/ilsvrc12_val_lmdb"
+    backend: LMDB
     batch_size: 50
   }
   transform_param {