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

Fix the ordering of attributes for indexlookup

Browse Source 
PiperOrigin-RevId: 294818441
Change-Id: I30e08ce32f9aedf1890b99b5e880e6aaa9cc9b8d
This commit is contained in:
Zhenyu Tan 2020-02-12 19:49:22 -08:00 committed by TensorFlower Gardener
parent 4d5741f1d4
commit 7fa9732b8d
1 changed files with 1 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
tensorflow/python/keras/layers/preprocessing/index_lookup.py
View File

@ -62,11 +62,11 @@ class IndexLookup(base_preprocessing_layer.CombinerPreprocessingLayer):
there is no cap on the size of the vocabulary. Note that the vocabulary
does include OOV buckets, so the effective number of unique values in the
vocabulary is `(max_tokens - num_oov_tokens)` when this value is set.
vocabulary: An optional list of vocabulary terms.
num_oov_tokens: The number of out-of-vocabulary tokens to use; defaults to
1. If this value is more than 1, OOV inputs are hashed to determine their
OOV value; if this value is 0, passing an OOV input will result in a
runtime error.
vocabulary: An optional list of vocabulary terms.
reserve_zero: Whether to reserve the index 0, which indicates pad values in
the Keras masking system. If True, the output of this layer will be in the
range `[1...max_tokens+1)`; if False, the output will be in the range