From 7251a4736daa795a1b6dca0ee5a76c88169e5e61 Mon Sep 17 00:00:00 2001 From: Julien Plu Date: Wed, 20 Jan 2021 15:04:53 +0100 Subject: [PATCH] Fix template (#9697) --- src/transformers/models/roberta/modeling_tf_roberta.py | 6 +++--- .../modeling_tf_{{cookiecutter.lowercase_modelname}}.py | 4 ++-- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/transformers/models/roberta/modeling_tf_roberta.py b/src/transformers/models/roberta/modeling_tf_roberta.py index 4df2cb2834..162aa2a197 100644 --- a/src/transformers/models/roberta/modeling_tf_roberta.py +++ b/src/transformers/models/roberta/modeling_tf_roberta.py @@ -307,7 +307,7 @@ class TFRobertaPooler(tf.keras.layers.Layer): return pooled_output -# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfAttention +# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfAttention with Bert->Roberta class TFRobertaSelfAttention(tf.keras.layers.Layer): def __init__(self, config, **kwargs): super().__init__(**kwargs) @@ -355,7 +355,7 @@ class TFRobertaSelfAttention(tf.keras.layers.Layer): attention_scores = tf.einsum("aecd,abcd->acbe", key_layer, query_layer) if attention_mask is not None: - # Apply the attention mask is (precomputed for all layers in TFBertModel call() function) + # Apply the attention mask is (precomputed for all layers in TFRobertaModel call() function) attention_scores = attention_scores + attention_mask # Normalize the attention scores to probabilities. @@ -375,7 +375,7 @@ class TFRobertaSelfAttention(tf.keras.layers.Layer): return outputs -# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfOutput with Bert->Roberta +# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfOutput class TFRobertaSelfOutput(tf.keras.layers.Layer): def __init__(self, config, **kwargs): super().__init__(**kwargs) diff --git a/templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_tf_{{cookiecutter.lowercase_modelname}}.py b/templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_tf_{{cookiecutter.lowercase_modelname}}.py index 1ae28aaaef..ce0cc3a63f 100644 --- a/templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_tf_{{cookiecutter.lowercase_modelname}}.py +++ b/templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/modeling_tf_{{cookiecutter.lowercase_modelname}}.py @@ -241,7 +241,7 @@ class TF{{cookiecutter.camelcase_modelname}}Embeddings(tf.keras.layers.Layer): -# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfAttention +# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfAttention with Bert->{{cookiecutter.camelcase_modelname}} class TF{{cookiecutter.camelcase_modelname}}SelfAttention(tf.keras.layers.Layer): def __init__(self, config, **kwargs): super().__init__(**kwargs) @@ -309,7 +309,7 @@ class TF{{cookiecutter.camelcase_modelname}}SelfAttention(tf.keras.layers.Layer) return outputs -# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfOutput with Bert->{{cookiecutter.camelcase_modelname}} +# Copied from transformers.models.bert.modeling_tf_bert.TFBertSelfOutput class TF{{cookiecutter.camelcase_modelname}}SelfOutput(tf.keras.layers.Layer): def __init__(self, config, **kwargs): super().__init__(**kwargs)