Question

我正在尝试将数据加载到this tutorial model进行测试。我基本上将标记的音频文件转换为图形文件，以便通过CNN进行分类。但我不确定邮票以及数据如何存储在字典中

广告输入我有一个列表： file1.wav classlabel file2.wav classlabel file3.wav classlabel

加载CNN分类器（遗憾的是我没有要检查的示例.pkl文件）：

dataset = pickle.load(open(os.path.join('data', prefix+'.pkl'), 'rb'))

def batch_input_fn(dataset, indices, batch_size=100, seed=None, num_epochs=1):  
    # Get the data into tensorflow
    stamps = np.array( dataset['stamp'] )[indices]
    print("stamps.shape:", stamps.shape)
    labels = np.array( dataset['label'] )[indices]
    print("labels.shape:", labels.shape)

我在这里很困惑，邮票到底代表什么？下面看来它既是阶级又是标签？

 def wav_to_stamp(prefix, word, wav):
    samples, sample_rate = soundfile.read( os.path.join('data', prefix, word, wav) )
    return samples_to_stamp(samples, sample_rate)


   for each wav file f that I have: 
        if not f.endswith('.wav'): continue
        #print(stamp_file)
        stamp = wav_to_stamp(prefix, word, stamp_file)
        # stamps are essentially images here created with stamp = scipy.misc.imresize(data_min0, (64, 32), 'bilinear')

   data_dictionary = dict(
    stamp=stamps, label=labels, 
   )

ds_file = os.path.join('data', prefix+save_as)
pickle.dump(data_dictionary, open(ds_file, 'wb'), protocol=pickle.HIGHEST_PROTOCOL)

上面的内容已经过一些修改以适应我的数据（read-in script中没有单词和句子。但我不确定邮票和标签是否合适。我如何正确地将它们存储在data_dictionary中，这样每张邮票都有标签？

Tensorflow CNN - 如何正确构建耦合到图像标记的标签字典

0 个答案: