繼續(xù)前面的三篇文章Part-1之拨，Part-2茉继，Part-3，這一篇我們來完善存儲和恢復(fù)機(jī)制蚀乔。

把計算圖保存到文件save_graph_to_file

下面是增加的代碼烁竭，先不要運行，稍后一起測試：

#將圖保存到文件,必要時創(chuàng)建允許的量子化    
def save_graph_to_file(graph, graph_file_name, module_spec, class_count):
    sess, _, _, _, _, _ = build_eval_session(module_spec, class_count)
    graph = sess.graph

    output_graph_def = tf.graph_util.convert_variables_to_constants(
        sess, graph.as_graph_def(), ['final_tensor_name'])

    with tf.gfile.FastGFile(graph_file_name, 'wb') as f:
        f.write(output_graph_def.SerializeToString())  

保存評估模型export_model

注意每次使用前必須把舊的saved_model文件夾刪除或改名吉挣。

#導(dǎo)出評估eval圖的模型pd文件用于提供服務(wù)
saved_model_dir=os.path.join(dir_path,'saved_model'+str(datetime.now()))
def export_model(module_spec, class_count):
    sess, in_image, _, _, _, _ = build_eval_session(module_spec, class_count)
    graph = sess.graph
    with graph.as_default():
        #輸入輸出點
        inputs = {'image': tf.saved_model.utils.build_tensor_info(in_image)}
        out_classes = sess.graph.get_tensor_by_name('final_tensor_name:0')
        outputs = {
            'prediction': tf.saved_model.utils.build_tensor_info(out_classes)
        }
        #創(chuàng)建簽名
        signature = tf.saved_model.signature_def_utils.build_signature_def(
            inputs=inputs,
            outputs=outputs,
            method_name=tf.saved_model.signature_constants.PREDICT_METHOD_NAME)

        #初始化
        legacy_init_op = tf.group(tf.tables_initializer(), name='legacy_init_op')

        #保存saved_model
        builder = tf.saved_model.builder.SavedModelBuilder(saved_model_dir)
        builder.add_meta_graph_and_variables(
            sess, [tf.saved_model.tag_constants.SERVING],
            signature_def_map={
                tf.saved_model.signature_constants.DEFAULT_SERVING_SIGNATURE_DEF_KEY:signature
            },
            legacy_init_op=legacy_init_op)
    builder.save()

改進(jìn)最終再訓(xùn)練函數(shù)

最后我們把run_final_retrain函數(shù)改進(jìn)一下派撕，增加評估eval和保存、導(dǎo)出功能睬魂。

注意每次使用前必須把舊的saved_model文件夾刪除或改名终吼。

以下是修改后的代碼（這里參數(shù)train_steps=10，所以得到的模型精度也非常糟糕氯哮。如果您的計算機(jī)允許际跪，官方默認(rèn)是4000，請量力而為）：

#保存概要和checkpoint路徑設(shè)置
CHECKPOINT_NAME = os.path.join(dir_path,'checkpoints/retrain')
summaries_dir=os.path.join(dir_path,'summaries/train')
ensure_dir_exists(os.path.join(dir_path,'output')) 
saved_model_path=os.path.join(dir_path,'output/out_graph.pd')
output_label_path=os.path.join(dir_path,'output/labels.txt')

#執(zhí)行訓(xùn)練兵保存checkpoint的函數(shù)
def run_final_retrain(train_steps=10,
             eval_step_interval=5,
             do_distort=True):
    module_spec = hub.load_module_spec(HUB_MODULE)
    
    #創(chuàng)建圖并獲取相關(guān)的張量入口
    graph, bottleneck_tensor, resized_image_tensor, wants_quantization = (
        create_module_graph(module_spec))
    
    with graph.as_default(): 
        #添加訓(xùn)練相關(guān)的張量和操作節(jié)點入口
        (train_step, cross_entropy, bottleneck_input,ground_truth_input,
         final_tensor) = add_final_retrain_ops(5, 'final_tensor_name', 
                                               bottleneck_tensor,wants_quantization,True)    

    with tf.Session(graph=graph) as sess:
        init = tf.global_variables_initializer()
        sess.run(init)
        
        #添加圖片解碼相關(guān)的張量入口操作
        jpeg_data_tensor, decoded_image_tensor = add_jpeg_decoding(module_spec)
        
        #讀取圖片的bottleneck數(shù)據(jù)
        if do_distort:
            distorted_jpeg_data_tensor,distorted_image_tensor=add_input_distortions(module_spec,True,50,50,50)
        else:
            cache_bottlenecks(sess, 
                              jpeg_data_tensor,decoded_image_tensor, 
                              resized_image_tensor,bottleneck_tensor)           
            
        #創(chuàng)建評估新層精度的操作
        evaluation_step, _ = add_evaluation_step(final_tensor, ground_truth_input)
        
        #記錄概要信息與保存
        train_saver = tf.train.Saver()
        merged = tf.summary.merge_all()
        train_writer = tf.summary.FileWriter(summaries_dir+'/retrain',sess.graph)      
        validation_writer = tf.summary.FileWriter(summaries_dir + '/validation')        
        
        #開始運作！
        for i in range(train_steps):
            #獲取圖片bottleneck數(shù)據(jù)
            if do_distort:
                (train_bottlenecks,train_ground_truth) = get_random_distorted_bottlenecks(
                    sess,BATCH_SIZE, 'training',
                    distorted_jpeg_data_tensor,distorted_image_tensor, 
                    resized_image_tensor, bottleneck_tensor)
            else:
                (train_bottlenecks,train_ground_truth, _) = get_random_cached_bottlenecks(
                    sess, BATCH_SIZE, 'training',
                    jpeg_data_tensor,decoded_image_tensor, 
                    resized_image_tensor, bottleneck_tensor)
            
            #啟動訓(xùn)練
            train_summary, _ = sess.run(
                [merged, train_step],
                feed_dict={bottleneck_input: train_bottlenecks,
                           ground_truth_input: train_ground_truth})
            train_writer.add_summary(train_summary, i)
            
            #間隔性啟動評估
            is_last_step = (i + 1 == train_steps)
            if (i % eval_step_interval) == 0 or is_last_step:
                train_accuracy, cross_entropy_value = sess.run(
                    [evaluation_step, cross_entropy],
                    feed_dict={bottleneck_input: train_bottlenecks,
                               ground_truth_input: train_ground_truth})
                
                tf.logging.info('%s: Step %d: Train accuracy = %.1f%%' %(datetime.now(), i, train_accuracy * 100))
                tf.logging.info('%s: Step %d: Cross entropy = %f' %(datetime.now(), i, cross_entropy_value))
            
                #使用不同的bottleneck數(shù)據(jù)進(jìn)行評估
                validation_bottlenecks, validation_ground_truth, _ = (
                    get_random_cached_bottlenecks(
                        sess, 10, 'validation', 
                        jpeg_data_tensor,decoded_image_tensor, 
                        resized_image_tensor, bottleneck_tensor))
                #啟動評估姆打！
                validation_summary, validation_accuracy = sess.run(
                    [merged, evaluation_step],
                    feed_dict={bottleneck_input: validation_bottlenecks,
                               ground_truth_input: validation_ground_truth})
                
                validation_writer.add_summary(validation_summary, i)
                tf.logging.info('%s: Step %d: Validation accuracy = %.1f%% (N=%d)' %(datetime.now(), i, validation_accuracy * 100,len(validation_bottlenecks)))            
            
        
            #間隔保存中介媒體文件良姆，為訓(xùn)練保存checkpoint
            if (i % eval_step_interval == 0 and i > 0):
                train_saver.save(sess, CHECKPOINT_NAME)
                intermediate_file_name = (os.path.join(dir_path + 'intermediate') + str(i) + '.pb')
                tf.logging.info('Save intermediate result to : '+intermediate_file_name)
                save_graph_to_file(graph, intermediate_file_name, module_spec,5)

        #保存模型    
        train_saver.save(sess, CHECKPOINT_NAME)        
        
        #執(zhí)行最終評估
        run_final_eval(sess, module_spec, 5,
                       jpeg_data_tensor, decoded_image_tensor,
                       resized_image_tensor,bottleneck_tensor)

        
        tf.logging.info('Save final result to : ' + saved_model_path)
        if wants_quantization:
            tf.logging.info('The model is instrumented for quantization with TF-Lite')
        save_graph_to_file(graph, saved_model_path, module_spec, 5)
        with tf.gfile.FastGFile(output_label_path, 'w') as f:
            f.write('\n'.join(image_lists.keys()) + '\n')
        export_model(module_spec, 5)

案例小結(jié)

這個案例來自Tensorflow官方教程，之前兩個相對都比較簡單幔戏，代碼量只有100行左右玛追，這個案例官方原代碼突然有1300行之多，大有才學(xué)了十以內(nèi)加減法然后就講微積分方程的感覺闲延。

這里整個案例去掉了很多官方代碼中我認(rèn)為無關(guān)緊要的部分痊剖，仍然有600多行，如果有時間我還會在整理這個案例垒玲，希望能只保留關(guān)鍵流程代碼陆馁，兩三百行不能再多了。

已經(jīng)讀到這里的用戶實屬難得侍匙，如果遇到困難氮惯，請從百度網(wǎng)盤下載(密碼:lzjg)直接下載final.py文件使用。請注意文件讀寫權(quán)限想暗，每次運行前請刪除saved_model文件夾。

探索人工智能的新邊界

如果您發(fā)現(xiàn)文章錯誤帘不，請不吝留言指正说莫；
如果您覺得有用，請點喜歡寞焙；
如果您覺得很有用储狭，感謝轉(zhuǎn)發(fā)~

END