2

我已经查找了此问题的先前答案,但尚未解决。我正在从头开始实施 YOLO 算法(用于对象检测)并且在训练部分遇到问题。

对于训练,我是 tf.estimator API,并且在 tensorflow示例中使用类似于 CNN MNIST 代码的代码。我收到以下错误:

Traceback (most recent call last):
  File "recover_v3.py", line 663, in <module>
    model.train(input_fn=train_input_fn, steps=1)
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 376, in train
    loss = self._train_model(input_fn, hooks, saving_listeners)
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 1145, in _train_model
    return self._train_model_default(input_fn, hooks, saving_listeners)
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 1170, in _train_model_default
    features, labels, model_fn_lib.ModeKeys.TRAIN, self.config)
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/estimator/estimator.py", line 1133, in _call_model_fn
    model_fn_results = self._model_fn(features=features, **kwargs)
  File "recover_v3.py", line 584, in cnn_model_fn
    loss=loss, global_step=tf.train.get_global_step())
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/training/optimizer.py", line 400, in minimize
    grad_loss=grad_loss)
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/training/optimizer.py", line 494, in compute_gradients
    self._assert_valid_dtypes([loss])
  File "/home/nyu-mmvc-019/miniconda3/envs/tf_0/lib/python3.6/site-packages/tensorflow/python/training/optimizer.py", line 872, in _assert_valid_dtypes
    dtype = t.dtype.base_dtype
AttributeError: 'NoneType' object has no attribute 'dtype'

主文件中损失函数相关代码如图所示(类似CNN MNIST官方示例):

if mode == tf.estimator.ModeKeys.TRAIN:
    # This gives the LOSS for each image in the batch.
    # It is importing loss function from another file (called loss_fn)
    # Apparently it returns None (not sure)
    loss = loss_fn.loss_fn(logits, labels)

    optimizer = tf.train.AdamOptimizer(learning_rate=params["learning_rate"])

    train_op = optimizer.minimize(
        loss=loss, global_step=tf.train.get_global_step())

    # Wrap all of this in an EstimatorSpec.
    spec = tf.estimator.EstimatorSpec(
        mode=mode,
        loss=loss,
        train_op=train_op,
        eval_metric_ops=None)

    return spec

以前对类似问题的回答表明损失函数没有返回任何内容。但是,当我尝试使用随机生成的数组的损失函数时,它可以正常工作并产生正常值。

另外,如果我从损失函数返回一个像 10.0 这样的常量,我仍然会得到同样的错误。

我不知道现在如何进行。另外,有什么方法可以打印损失函数返回的损失。显然,tf.estimator API 自己启动一个 tensorflow 会话,如果我尝试创建另一个会话(为了打印损失函数返回的值),我会得到其他错误。

4

1 回答 1

0

但是,当我尝试使用随机生成的数组的损失函数时,它可以正常工作并产生正常值。

您的 input_fn 似乎有问题。您确定它已正确实施吗?

另外,有什么方法可以打印损失函数返回的损失。

Estimator 每隔 global_step % 'save_summary_steps' 在控制台中自动打印损失函数的值。您还可以使用标量摘要跟踪损失函数,如下所示:

tf.summary.scalar('loss', loss)
于 2018-09-21T16:07:28.027 回答