Why my feature extraction script fails ?

[jaba marwen]

Hi every one ,

I have used the following script , to extract features vectors of my dataset using AlexNet pretrained model.

import numpy as np
import hickle as hkl
import caffe


caffe.set_mode_gpu()

def feature_extract(img):

    model_file='/home/jaba/caffe/data/diota_model/feature_extractor/bvlc_reference_caffenet.caffemodel'
    deploy_file='/home/jaba/caffe/data/diota_model/feature_extractor/alex.deployprototxt'

    net=caffe.Net(deploy_file,model_file,caffe.TEST)

    mean_values=np.array([103.939, 116.779, 123.68])

    #setting the transformer
   
    transformer = caffe.io.Transformer({'data': net.blobs['data'].data.shape})
    transformer.set_mean('data', mean_values ) #subtract by mean
    transformer.set_transpose('data', (2,0,1)) #(H,W,C) => (C,H,W)
    transformer.set_raw_scale('data', 255.0)   #[0.0, 1.0] => [0.0, 255.0]
    transformer.set_channel_swap("data", (2,1,0)) # RGB => BGR

    #img = caffe.io.load_image(img)

    net.blobs['data'].data[...] = transformer.preprocess('data', img)
   
    output=net.forward()
   
    feat=net.blobs['fc7'].data.copy()

    return feat





def create_dataset(datalist,db_prefix):
    with open(datalist) as fr:
            lines = fr.readlines()
    lines = [line.rstrip() for line in lines]

    feats = []
        labels = []
   
    for line_i, line in enumerate(lines):

            a=len(line)
        label=line[a-1]
        img_path=line[0:a-2]
            img = caffe.io.load_image(img_path)
            feat = feature_extract(img)
            feats.append(feat)
            label = int(label)
            labels.append(label)
            if (line_i + 1) % 100 == 0:
                    print "processed", line_i + 1


    feats = np.asarray(feats)
        labels = np.asarray(labels)


    hkl.dump(feats, dbprefix + "_features.hkl", mode="w")
        hkl.dump(labels, dbprefix + "_labels.hkl", mode="w")




create_dataset('train.txt','vgg_fc7_train')
create_dataset('val.txt','vgg_fc7_test')

 I have got the following error after few iterations error == cudaSuccess (2 vs. 0)  out of memory .
 I know that's my gpu is out of memory . Would you help me to correct this issue ?
 For train .tx and val.txt files , train.txt ( the path of 5000 images ) and val.txt (the path of 1000 images ) .
 
Thank you .