Variable conv1/weights already exists, disallowed. Did you mean to set reuse=True in VarScope? 解决方法

转载自： https://www.jianshu.com/p/fc11f32800f9

 

保存checkpoint后，尝试调用保存的模型：

 

 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')

上面是Tensorflow官网给出的模型调用方法，但出现了下面的错误：

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

原因是在一次测试过程中，测试了多张图片，导致模型的参数需要重复使用，所以需要告诉TF‘允许复用参数’，所以只要在上面的代码加上

tf.get_variable_scope().reuse_variables()

Error就会消失。

 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')

转载自： https://www.jianshu.com/p/fc11f32800f9