diff options
| author |  Lukasz Kaiser <lukaszkaiser@google.com> | 2016-10-14 11:51:28 -0800 |
|---|---|---|
| committer |  TensorFlower Gardener <gardener@tensorflow.org> | 2016-10-14 13:04:42 -0700 |
| commit | 3a8ec777219002edf05264730f06760d1c346d6f (patch) | |
| tree | 61d20de074a33db9abf160af9a6a34f435850d06 | |
| parent | 68af6c9848defff9c9e77bf92302fd266f019cd5 (diff) | |
Adding a comment to seq2seq to warn about large size of attention
when no output_projection is specified (see #4938).
Change: 136187392
| -rw-r--r-- | tensorflow/python/ops/seq2seq.py | 3 |
1 files changed, 3 insertions, 0 deletions
diff --git a/tensorflow/python/ops/seq2seq.py b/tensorflow/python/ops/seq2seq.py index 7a4b547fac..9ec12583de 100644 --- a/tensorflow/python/ops/seq2seq.py +++ b/tensorflow/python/ops/seq2seq.py @@ -772,6 +772,9 @@ def embedding_attention_seq2seq(encoder_inputs, input_size]). Then it runs attention decoder, initialized with the last encoder state, on embedded decoder_inputs and attending to encoder outputs. + Warning: when output_projection is None, the size of the attention vectors + and variables will be made proportional to num_decoder_symbols, can be large. + Args: encoder_inputs: A list of 1D int32 Tensors of shape [batch_size]. decoder_inputs: A list of 1D int32 Tensors of shape [batch_size]. |