bert4keras
A Keras/TensorFlow-based implementation library for BERT-series models, developed by Su Jianlin. Known for its minimalist coding style, it has gained wide adoption in the Chinese NLP community.
Timeline (last 90 days)
The design philosophy of bert4torch draws inspiration from bert4keras in the Keras ecosystem.
bert4keras was developed by Su Jianlin (Su Shen) and is based on the Keras/TensorFlow ecosystem.
bert4keras implements models including BERT, RoBERTa, ALBERT, and GPT series with minimalist code.
bert4torch intentionally aligns with bert4keras usage conventions in API naming, model loading methods, and training loop design.
All Facts (4)
The design philosophy of bert4torch draws inspiration from bert4keras in the Keras ecosystem.
50%Unverifiedbert4keras was developed by Su Jianlin (Su Shen) and is based on the Keras/TensorFlow ecosystem.
50%Unverifiedbert4keras implements models including BERT, RoBERTa, ALBERT, and GPT series with minimalist code.
50%Unverifiedbert4torch intentionally aligns with bert4keras usage conventions in API naming, model loading methods, and training loop design.
50%