UER-py is a framework for pre-training models (implemented by PyTorch). UER-py facilitates the implementation of existing pre-training models (e.g. BERT), and supports further improvements. UER-py also open-sources a series of Chinese pre-training models based on different corpora, encoders, and targets. With these open-source pre-training models, we can obtain significant improvements compared with Google BERT, achieving new SOTA results on a range of Chinese datasets.
The address is https://github.com/dbiir/UER-py
I hope you could add this to the list, thank you very much~
Jiakui
commented
5 years ago
UER-py is a framework for pre-training models (implemented by PyTorch). UER-py facilitates the implementation of existing pre-training models (e.g. BERT), and supports further improvements. UER-py also open-sources a series of Chinese pre-training models based on different corpora, encoders, and targets. With these open-source pre-training models, we can obtain significant improvements compared with Google BERT, achieving new SOTA results on a range of Chinese datasets. The address is https://github.com/dbiir/UER-py I hope you could add this to the list, thank you very much~