The recipe focuses on building an image classification model using Transfer Learning. It will utilize the dataset prepared in the previous recipes and use the Inception-BN architecture. The BN in Inception-BN stands for batch normalization. Details of the Inception model in computer vision can be found in Szegedy et al. (2015).
The section cover's the prerequisite to set-up a classification model using INCEPTION-BN pretrained model.
- Convert images into
.rec
file for train and validation. - Download the Inception-BN architecture from http://data.dmlc.ml/models/imagenet/inception-bn/.
- Install R and the
mxnet
package in R.
- Load the
.rec
file as iterators. The following is the function to load the.rec
data as iterators:
# Function to load data as iterators data.iterator <- function(data.shape, train.data, val.data, BATCHSIZE = 128) { # Load training data as iterator train <- mx.io.ImageRecordIter( path.imgrec ...