Using google bert for RoBERTa

first, using convert_pytorch_checkpoint_to_tf.py convert wieght

second, edit config file

third, modified modelling:using_token_type=False start=2

Fourth, change bert tokenizer to roberta tokenizer

Name		Name	Last commit message	Last commit date
Latest commit History 16 Commits
.idea		.idea
.gitignore		.gitignore
CONTRIBUTING.md		CONTRIBUTING.md
LICENSE		LICENSE
README.md		README.md
__init__.py		__init__.py
bash.py		bash.py
configuration_bert.py		configuration_bert.py
configuration_roberta.py		configuration_roberta.py
configuration_utils.py		configuration_utils.py
convert_pytorch_checkpoint_to_tf.py		convert_pytorch_checkpoint_to_tf.py
create_pretraining_data.py		create_pretraining_data.py
eval_roberta.py		eval_roberta.py
eval_roberta_align.py		eval_roberta_align.py
extract_features.py		extract_features.py
file_utils.py		file_utils.py
modeling.py		modeling.py
modeling_align.py		modeling_align.py
modeling_bert.py		modeling_bert.py
modeling_roberta.py		modeling_roberta.py
modeling_test.py		modeling_test.py
modeling_utils.py		modeling_utils.py
multilingual.md		multilingual.md
optimization.py		optimization.py
optimization_align.py		optimization_align.py
optimization_test.py		optimization_test.py
predicting_movie_reviews_with_bert_on_tf_hub.ipynb		predicting_movie_reviews_with_bert_on_tf_hub.ipynb
requirements.txt		requirements.txt
run_classifier.py		run_classifier.py
run_classifier_align.py		run_classifier_align.py
run_classifier_with_tfhub.py		run_classifier_with_tfhub.py
run_pretraining.py		run_pretraining.py
run_squad.py		run_squad.py
sample_text.txt		sample_text.txt
tokenization.py		tokenization.py
tokenization_gpt2.py		tokenization_gpt2.py
tokenization_roberta.py		tokenization_roberta.py
tokenization_test.py		tokenization_test.py
tokenization_utils.py		tokenization_utils.py

Provide feedback