Skip to content

GitLab

Commit a8ac58ea authored by Hongkun Yu's avatar Hongkun Yu Committed by A. Unique TensorFlower
Browse files

Update comments and users should use BertPretrainerV2 in the future. 

PiperOrigin-RevId: 344194437
parent d2aac5d7
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 5 deletions
official/nlp/modeling/models/bert_pretrainer.py
View file @ a8ac58ea
...@@ -28,11 +28,9 @@ from official.nlp.modeling import networks ...@@ -28,11 +28,9 @@ from official.nlp.modeling import networks
@tf.keras.utils.register_keras_serializable(package='Text') @tf.keras.utils.register_keras_serializable(package='Text')
class BertPretrainer(tf.keras.Model): class BertPretrainer(tf.keras.Model):
"""BERT network training model. """BERT pretraining model.
This is an implementation of the network structure surrounding a transformer [Note] Please use the new BertPretrainerV2 for your projects.
encoder as described in "BERT: Pre-training of Deep Bidirectional Transformers
for Language Understanding" (https://arxiv.org/abs/1810.04805).
The BertPretrainer allows a user to pass in a transformer stack, and The BertPretrainer allows a user to pass in a transformer stack, and
instantiates the masked language model and classification networks that are instantiates the masked language model and classification networks that are
...@@ -159,7 +157,6 @@ class BertPretrainer(tf.keras.Model): ...@@ -159,7 +157,6 @@ class BertPretrainer(tf.keras.Model):
return cls(**config) return cls(**config)
# TODO(hongkuny): Migrate to BertPretrainerV2 for all usages.
@tf.keras.utils.register_keras_serializable(package='Text') @tf.keras.utils.register_keras_serializable(package='Text')
@gin.configurable @gin.configurable
class BertPretrainerV2(tf.keras.Model): class BertPretrainerV2(tf.keras.Model):
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment