experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

[tf contrib seq2seq] Provide informative error messages in AttentionWrapper.call.

Browse Source 
Related issues: #11077, #11540.

PiperOrigin-RevId: 162362081
This commit is contained in:
Eugene Brevdo 2017-07-18 09:30:16 -07:00 committed by TensorFlower Gardener
parent 235060d7e7
commit a3d3ce1603
1 changed files with 7 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
tensorflow/contrib/seq2seq/python/ops/attention_wrapper.py
View File

@ -1137,7 +1137,14 @@ class AttentionWrapper(rnn_cell_impl.RNNCell):
- `attention_or_cell_output` depending on `output_attention`. - `attention_or_cell_output` depending on `output_attention`.
- `next_state` is an instance of `AttentionWrapperState` - `next_state` is an instance of `AttentionWrapperState`
containing the state calculated at this time step. containing the state calculated at this time step.
Raises:
TypeError: If `state` is not an instance of `AttentionWrapperState`.
""" """
if not isinstance(state, AttentionWrapperState):
raise TypeError("Expected state to be instance of AttentionWrapperState. "
"Received type %s instead." % type(state))
# Step 1: Calculate the true inputs to the cell based on the # Step 1: Calculate the true inputs to the cell based on the
# previous attention value. # previous attention value.
cell_inputs = self._cell_input_fn(inputs, state.attention) cell_inputs = self._cell_input_fn(inputs, state.attention)