System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow):
  Yes.
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04):
  Linux Ubuntu 18
• TensorFlow installed from (source or binary):
  Binary
• TensorFlow version (use command below):
  v1.8.0-3410-g79d7e11f3e, 1.10.0-dev20180616
• Python version:
  3.6
• Bazel version (if compiling from source):
  N/A
• GCC/Compiler version (if compiling from source):
  N/A
• CUDA/cuDNN version:
  CUDA 9.0, cuDNN 7.0
• GPU model and memory:
  NVIDIA GeForce 1080
• Exact command to reproduce:
  Running the provided script should produce the error. The attached file shows the output.

Describe the problem

It's not possible to debug Keras models since BaseDebugWrapperSession in tensorflow/python/debug/wrappers/framework.py does not implement the _make_callable_from_options method.

Source code / logs

• Script:

import tensorflow as tf
import numpy as np
from tensorflow import keras
from tensorflow.python import debug as tf_debug


inputs = np.zeros((10, 3))
targets = np.zeros((10, 4))
dataset = tf.data.Dataset.from_tensor_slices((inputs, targets))
dataset = dataset.repeat(100)
dataset = dataset.batch(5)

with tf_debug.LocalCLIDebugWrapperSession(tf.Session()) as sess:
    keras.backend.set_session(sess)

    x = keras.layers.Input(shape=(3,), name='input')
    flat = keras.layers.Flatten()(x)
    y = keras.layers.Dense(4, name='dense')(flat)

    model = keras.Model(x, y)
    model.compile(loss='mse', optimizer='rmsprop')

    model.fit(dataset, epochs=1, steps_per_epoch=2, verbose=1)

• Output:
  tfdbg_output.txt