Keras与TensorBoard错误“InvalidArgumentError：您必须养活一个值占位符张量”

Question

我试图让TensorBoard与Keras工作。它看起来像我能够与TF（1.12.0）和keras（2.1.6-TF）运行的初始模型。我有一些简单的代码。下面列出：

%matplotlib inline
from io import StringIO

import numpy as np
import pandas as pd
import tensorflow as tf

csv = StringIO('''a,b,c,y
0,1,2,0
1,2,0,1
0,2,1,0
3,2,1,1
3,1,2,0''')
data = pd.read_csv(csv)

def tb_cb(batch_size):
    # visualize graphs and grandient
    tb = tf.keras.callbacks.TensorBoard(log_dir='/tmp/test/',
                                       histogram_freq=1,
                                       batch_size=batch_size, write_graph=True,
                                       write_grads=True)
    return tb

m = tf.keras.Sequential([
   # going to change 1 in the line below
    tf.keras.layers.Dense(1, activation='relu', input_shape=(3,), name='hidden1'),
    tf.keras.layers.Dense(1, activation='linear', name='output')
])
m.compile(loss='mse', optimizer='adam', metrics=['mae'])
X = data.iloc[:,:3]
y = data.y
hist = m.fit(X, y, epochs=10, verbose=1, callbacks=[tb_cb(10)],
            validation_data=(X,y))

我第一次运行此我得到TensorBoard输出。然后，我已经改变神经元的数目隐藏层，并重新运行模型。

我得到以下错误：

---------------------------------------------------------------------------
InvalidArgumentError                      Traceback (most recent call last)
<ipython-input-11-4e4fb6f60bf0> in <module>
     15 y = data.y
     16 hist = m.fit(X, y, epochs=10, verbose=1, callbacks=[tb_cb(10)],
---> 17             validation_data=(X,y))

~/.env/364/lib/python3.6/site-packages/tensorflow/python/keras/engine/training.py in fit(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, max_queue_size, workers, use_multiprocessing, **kwargs)
   1637           initial_epoch=initial_epoch,
   1638           steps_per_epoch=steps_per_epoch,
-> 1639           validation_steps=validation_steps)
   1640 
   1641   def evaluate(self,

~/.env/364/lib/python3.6/site-packages/tensorflow/python/keras/engine/training_arrays.py in fit_loop(model, inputs, targets, sample_weights, batch_size, epochs, verbose, callbacks, val_inputs, val_targets, val_sample_weights, shuffle, initial_epoch, steps_per_epoch, validation_steps)
    231                 sample_weights=val_sample_weights,
    232                 batch_size=batch_size,
--> 233                 verbose=0)
    234             if not isinstance(val_outs, list):
    235               val_outs = [val_outs]

~/.env/364/lib/python3.6/site-packages/tensorflow/python/keras/engine/training_arrays.py in test_loop(model, inputs, targets, sample_weights, batch_size, verbose, steps)
    437         ins_batch[i] = ins_batch[i].toarray()
    438 
--> 439       batch_outs = f(ins_batch)
    440 
    441       if isinstance(batch_outs, list):

~/.env/364/lib/python3.6/site-packages/tensorflow/python/keras/backend.py in __call__(self, inputs)
   2984 
   2985     fetched = self._callable_fn(*array_vals,
-> 2986                                 run_metadata=self.run_metadata)
   2987     self._call_fetch_callbacks(fetched[-len(self._fetches):])
   2988     return fetched[:len(self.outputs)]

~/.env/364/lib/python3.6/site-packages/tensorflow/python/client/session.py in __call__(self, *args, **kwargs)
   1437           ret = tf_session.TF_SessionRunCallable(
   1438               self._session._session, self._handle, args, status,
-> 1439               run_metadata_ptr)
   1440         if run_metadata:
   1441           proto_data = tf_session.TF_GetBuffer(run_metadata_ptr)

~/.env/364/lib/python3.6/site-packages/tensorflow/python/framework/errors_impl.py in __exit__(self, type_arg, value_arg, traceback_arg)
    526             None, None,
    527             compat.as_text(c_api.TF_Message(self.status.status)),
--> 528             c_api.TF_GetCode(self.status.status))
    529     # Delete the underlying status object from memory otherwise it stays alive
    530     # as there is a reference to status from this from the traceback due to

InvalidArgumentError: You must feed a value for placeholder tensor 'dense_9_target' with dtype float and shape [?,?]
     [[{{node dense_9_target}} = Placeholder[dtype=DT_FLOAT, shape=[?,?], _device="/job:localhost/replica:0/task:0/device:CPU:0"]()]]

Answer 1

你总是应该构建模型，以避免在您的TF图表上一次运行的剩余节点之前调用Keras.clear_session（）。

所以加keras.backend.clear_session（）创建模型权利之前。