def predict(models_path,image_dir,labels_filename,labels_nums, data_format): [batch_size, resize_height, resize_width, depths] = data_format #labels = np.loadtxt(labels_filename, str, delimiter='\t') input_images = tf.placeholder(dtype=tf.float32, shape=[None, resize_height, resize_width, depths], name='input') #其他模型预测请修改这里 with slim.arg_scope(inception_v1.inception_v1_arg_scope()): out, end_points = inception_v1.inception_v1(inputs=input_images, num_classes=labels_nums, dropout_keep_prob=1.0, is_training=False) # 将输出结果进行softmax分布,再求最大概率所属类别 score = tf.nn.softmax(out,name='pre') class_id = tf.argmax(score, 1) sess = tf.InteractiveSession() sess.run(tf.global_variables_initializer()) saver = tf.train.Saver() saver.restore(sess, models_path) images_list=glob.glob(os.path.join(image_dir,'*.jpg')) for image_path in images_list: im=read_image(image_path,resize_height,resize_width,normalization=True) im=im[np.newaxis,:] #pred = sess.run(f_cls, feed_dict={x:im, keep_prob:1.0}) pre_score,pre_label = sess.run([score,class_id], feed_dict={input_images:im}) max_score=pre_score[0,pre_label] print("{} is: pre labels:{},name:{} score: {}".format(image_path,pre_label,list(labels_filename.keys())[list(labels_filename.values()).index(pre_label)], max_score)) sess.close()
def testModelHasExpectedNumberOfParameters(self): batch_size = 5 height, width = 224, 224 inputs = random_ops.random_uniform((batch_size, height, width, 3)) with arg_scope(inception_v1.inception_v1_arg_scope()): inception_v1.inception_v1_base(inputs) total_params, _ = model_analyzer.analyze_vars( variables_lib.get_model_variables()) self.assertAlmostEqual(5607184, total_params)
def predict_images(): models_path = './logs/model.ckpt-11999' images_dir = './test_images' labels_txt_file = './dataset/label.txt' num_calsses = 5 resize_height = 224 resize_width = 224 channels = 3 images_list = glob.glob(os.path.join(images_dir, '*.jpg')) # 返回匹配路径名模式的路径列表 # delimiter='\t'表示以空格隔开 labels = np.loadtxt( labels_txt_file, str, delimiter='\t' ) # labels = ['flower' 'guitar' 'animal' 'houses' 'plane'] intput_images = tf.placeholder( dtype=tf.float32, shape=[None, resize_height, resize_width, channels], name='input') with slim.arg_scope(inception_v1.inception_v1_arg_scope()): out, end_points = inception_v1.inception_v1(inputs=intput_images, num_classes=num_calsses, dropout_keep_prob=1.0, is_training=False) score = tf.nn.softmax(out) class_id = tf.argmax(score, axis=1) # 最大score的id值 init = tf.global_variables_initializer() with tf.Session() as sess: sess.run(init) saver = tf.train.Saver() saver.restore(sess, models_path) for image_name in images_list: image = read_image(image_name, resize_height, resize_width, normalization=True) image = image[np.newaxis, :] # 给数据增加一个新的维度 predict_score, predict_id = sess.run( [score, class_id], feed_dict={intput_images: image}) max_score = predict_score[0, predict_id] # id相对应的得分(得到的score是二维的) print("{} is: label:{},name:{} score: {}".format( image_name, predict_id, labels[predict_id], max_score))
def train(train_tfrecords_file, base_lr, max_steps, val_tfrecords_file, num_classes, data_shape, train_log_dir, val_nums): """ :param train_tfrecords_file: 训练数据集的tfrecords文件 :param base_lr: 学习率 :param max_steps: 迭代次数 :param val_tfrecords_file: 验证数据集的tfrecords文件 :param num_classes: 分类个数 :param data_shape: 数据形状[batch_size, resize_height, resize_width, channels] :param train_log_dir: 模型文件的存放位置 :return: """ [batch_size, resize_height, resize_width, channels] = data_shape # 读取训练数据 train_images, train_labels = read_tfrecords(train_tfrecords_file, resize_height, resize_width, output_model='normalization') train_batch_images, train_batch_labels = get_batch_images( train_images, train_labels, batch_size=batch_size, num_classes=num_classes, one_hot=True, shuffle=True) # 读取验证数据,验证数据集可以不用打乱 val_images, val_labels = read_tfrecords(val_tfrecords_file, resize_height, resize_width, output_model='normalization') val_batch_images, val_batch_labels = get_batch_images( val_images, val_labels, batch_size=batch_size, num_classes=num_classes, one_hot=True, shuffle=False) with slim.arg_scope(inception_v1.inception_v1_arg_scope() ): # inception_v1.inception_v1_arg_scope()括号不能掉,表示一个函数 out, end_points = inception_v1.inception_v1( inputs=input_images, num_classes=num_classes, is_training=is_training, dropout_keep_prob=keep_prob) loss = tf.losses.softmax_cross_entropy(onehot_labels=input_labels, logits=out) accuracy = tf.reduce_mean( tf.cast(tf.equal(tf.argmax(out, 1), tf.argmax(input_labels, 1)), tf.float32)) * 100.0 optimizer = tf.train.MomentumOptimizer(learning_rate=base_lr, momentum=0.9) # 这里可以使用不同的优化函数 # 在定义训练的时候, 注意到我们使用了`batch_norm`层时,需要更新每一层的`average`和`variance`参数, # 正常的训练过程不包括更新,需要我们去手动像下面这样更新 with tf.control_dependencies(tf.get_collection( tf.GraphKeys.UPDATE_OPS)): # 执行完更新操作之后,再进行训练操作 train_op = slim.learning.create_train_op(total_loss=loss, optimizer=optimizer) saver = tf.train.Saver() init = tf.global_variables_initializer() with tf.Session() as sess: sess.run(init) coord = tf.train.Coordinator() threads = tf.train.start_queue_runners(sess=sess, coord=coord) for steps in np.arange(max_steps): input_batch_images, input_batch_labels = sess.run( [train_batch_images, train_batch_labels]) _, train_loss = sess.run( [train_op, loss], feed_dict={ input_images: input_batch_images, input_labels: input_batch_labels, keep_prob: 0.8, is_training: True }) # 得到训练过程中的loss, accuracy值 if steps % 50 == 0 or (steps + 1) == max_steps: train_acc = sess.run(accuracy, feed_dict={ input_images: input_batch_images, input_labels: input_batch_labels, keep_prob: 1.0, is_training: False }) print('Step: %d, loss: %.4f, accuracy: %.4f' % (steps, train_loss, train_acc)) # 在验证数据集上得到loss, accuracy值 if steps % 200 == 0 or (steps + 1) == max_steps: val_images_batch, val_labels_batch = sess.run( [val_batch_images, val_batch_labels]) val_loss, val_acc = sess.run( [loss, accuracy], feed_dict={ input_images: val_images_batch, input_labels: val_labels_batch, keep_prob: 1.0, is_training: False }) val_loss, val_acc = evaluation(sess, loss, accuracy, val_batch_images, val_batch_labels, val_nums) print( '** Step %d, val loss = %.2f, val accuracy = %.2f%% **' % (steps, val_loss, val_acc)) # 每隔2000步储存一下模型文件 if steps % 2000 == 0 or (steps + 1) == max_steps: checkpoint_path = os.path.join(train_log_dir, 'model.ckpt') saver.save(sess, checkpoint_path, global_step=steps) coord.request_stop() coord.join(threads)
def train(train_record_file, train_log_step, train_param, val_record_file, val_log_step, labels_nums, data_shape, snapshot, snapshot_prefix): ''' :param train_record_file: 训练的tfrecord文件 :param train_log_step: 显示训练过程log信息间隔 :param train_param: train参数 :param val_record_file: 验证的tfrecord文件 :param val_log_step: 显示验证过程log信息间隔 :param val_param: val参数 :param labels_nums: labels数 :param data_shape: 输入数据shape :param snapshot: 保存模型间隔 :param snapshot_prefix: 保存模型文件的前缀名 :return: ''' [base_lr,max_steps]=train_param [batch_size,resize_height,resize_width,depths]=data_shape # 获得训练和测试的样本数 train_nums=get_example_nums(train_record_file) val_nums=get_example_nums(val_record_file) print('train nums:%d,val nums:%d'%(train_nums,val_nums)) # 从record中读取图片和labels数据 # train数据,训练数据一般要求打乱顺序shuffle=True train_images, train_labels = read_records(train_record_file, resize_height, resize_width, type='normalization') train_images_batch, train_labels_batch = get_batch_images(train_images, train_labels, batch_size=batch_size, labels_nums=labels_nums, one_hot=True, shuffle=True) # val数据,验证数据可以不需要打乱数据 val_images, val_labels = read_records(val_record_file, resize_height, resize_width, type='normalization') val_images_batch, val_labels_batch = get_batch_images(val_images, val_labels, batch_size=batch_size, labels_nums=labels_nums, one_hot=True, shuffle=False) # Define the model: with slim.arg_scope(inception_v1.inception_v1_arg_scope()): out, end_points = inception_v1.inception_v1(inputs=input_images, num_classes=labels_nums, dropout_keep_prob=keep_prob, is_training=is_training) # Specify the loss function: tf.losses定义的loss函数都会自动添加到loss函数,不需要add_loss()了 tf.losses.softmax_cross_entropy(onehot_labels=input_labels, logits=out)#添加交叉熵损失loss=1.6 # slim.losses.add_loss(my_loss) loss = tf.losses.get_total_loss(add_regularization_losses=False)#添加正则化损失loss=2.2 accuracy = tf.reduce_mean(tf.cast(tf.equal(tf.argmax(out, 1), tf.argmax(input_labels, 1)), tf.float32)) # Specify the optimization scheme: # optimizer = tf.train.GradientDescentOptimizer(learning_rate=base_lr) # global_step = tf.Variable(0, trainable=False) # learning_rate = tf.train.exponential_decay(0.05, global_step, 150, 0.9) # optimizer = tf.train.MomentumOptimizer(learning_rate=base_lr,momentum= 0.9) # # train_tensor = optimizer.minimize(loss, global_step) # train_op = slim.learning.create_train_op(loss, optimizer,global_step=global_step) # 在定义训练的时候, 注意到我们使用了`batch_norm`层时,需要更新每一层的`average`和`variance`参数, # 更新的过程不包含在正常的训练过程中, 需要我们去手动像下面这样更新 # 通过`tf.get_collection`获得所有需要更新的`op` update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS) # 使用`tensorflow`的控制流, 先执行更新算子, 再执行训练 with tf.control_dependencies(update_ops): # create_train_op that ensures that when we evaluate it to get the loss, # the update_ops are done and the gradient updates are computed. # train_op = slim.learning.create_train_op(total_loss=loss,optimizer=optimizer) train_op = slim.learning.create_train_op(total_loss=loss, optimizer=optimizer) # 循环迭代过程 step_train(train_op, loss, accuracy, train_images_batch, train_labels_batch, train_nums, train_log_step, val_images_batch, val_labels_batch, val_nums, val_log_step, snapshot_prefix, snapshot)