愿天堂沒(méi)有Tensorflow！ 阿門(mén)。

ValueError: Variable conv1/weights already exists, disallowed. Did you mean to set reuse=True in VarScope? Originally defined at:

將一個(gè)模型訓(xùn)練好，并且保存了checkpoint后，我嘗試調(diào)用保存的模型：

**保存的模型文件**

 with tf.Session() as sess:                
                 print("Reading checkpoints...")
                 ckpt = tf.train.get_checkpoint_state(logs_train_dir)
                 if ckpt and ckpt.model_checkpoint_path:
                     global_step = ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
                     saver.restore(sess, ckpt.model_checkpoint_path)
                     print('Loading success, global_step is %s' % global_step)
                 else:
                     print('No checkpoint file found')

這是Tensorflow官網(wǎng)給出的模型調(diào)用方法，可是卻出現(xiàn)了下面的錯(cuò)誤：

ValueError: Variable conv1/weights already exists, disallowed. Did you mean to set reuse=True in VarScope? Originally defined at:

原因是我們?cè)谝淮螠y(cè)試過(guò)程中，測(cè)試了多張圖片，導(dǎo)致我們模型的參數(shù)需要重復(fù)使用，所以我們需要告訴TF‘允許復(fù)用參數(shù)’，所以只要在上面的代碼加上

tf.get_variable_scope().reuse_variables()

Error就會(huì)消失。

 with tf.Session() as sess:
                 tf.get_variable_scope().reuse_variables()         
                 print("Reading checkpoints...")
                 ckpt = tf.train.get_checkpoint_state(logs_train_dir)
                 if ckpt and ckpt.model_checkpoint_path:
                     global_step = ckpt.model_checkpoint_path.split('/')[-1].split('-')[-1]
                     saver.restore(sess, ckpt.model_checkpoint_path)
                     print('Loading success, global_step is %s' % global_step)
                 else:
                     print('No checkpoint file found')

**Error消失**