Model Overview
Model Features
Model Capabilities
Use Cases
🚀 Chinese word-based RoBERTa Miniatures
This project presents a set of 5 Chinese word-based RoBERTa models. These models offer faster processing and better performance compared to character-based models, thanks to their shorter sequence lengths. They are pre-trained using publicly available resources, and all training details are provided for easy reproduction.
🚀 Quick Start
You can use this model directly with a pipeline for masked language modeling. Here is an example using the word-based RoBERTa-Medium:
>>> from transformers import pipeline
>>> unmasker = pipeline('fill-mask', model='uer/roberta-medium-word-chinese-cluecorpussmall')
>>> unmasker("[MASK]的首都是北京。")
[
{'sequence': '中国 的首都是北京。',
'score': 0.21525809168815613,
'token': 2873,
'token_str': '中国'},
{'sequence': '北京 的首都是北京。',
'score': 0.15194718539714813,
'token': 9502,
'token_str': '北京'},
{'sequence': '我们 的首都是北京。',
'score': 0.08854265511035919,
'token': 4215,
'token_str': '我们'},
{'sequence': '美国 的首都是北京。',
'score': 0.06808705627918243,
'token': 7810,
'token_str': '美国'},
{'sequence': '日本 的首都是北京。',
'score': 0.06071401759982109,
'token': 7788,
'token_str': '日本'}
]
✨ Features
- Word-based: Compared with most Chinese pre-trained weights based on Chinese characters, word-based models are faster and have better performance according to experimental results.
- Multiple Sizes: Five different sizes of Chinese word-based RoBERTa models are released to meet various needs.
- Reproducibility: A publicly available corpus and word segmentation tool are used, and all training details are provided.
📦 Installation
There is no specific installation steps provided in the original README. If you want to use these models, you can directly download them from the UER-py Modelzoo page or via HuggingFace from the links below:
Link | |
---|---|
word-based RoBERTa-Tiny | L=2/H=128 (Tiny) |
word-based RoBERTa-Mini | L=4/H=256 (Mini) |
word-based RoBERTa-Small | L=4/H=512 (Small) |
word-based RoBERTa-Medium | L=8/H=512 (Medium) |
word-based RoBERTa-Base | L=12/H=768 (Base) |
💻 Usage Examples
Basic Usage
from transformers import AlbertTokenizer, BertModel
tokenizer = AlbertTokenizer.from_pretrained('uer/roberta-medium-word-chinese-cluecorpussmall')
model = BertModel.from_pretrained("uer/roberta-medium-word-chinese-cluecorpussmall")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='pt')
output = model(**encoded_input)
Advanced Usage
from transformers import AlbertTokenizer, TFBertModel
tokenizer = AlbertTokenizer.from_pretrained('uer/roberta-medium-word-chinese-cluecorpussmall')
model = TFBertModel.from_pretrained("uer/roberta-medium-word-chinese-cluecorpussmall")
text = "用你喜欢的任何文本替换我。"
encoded_input = tokenizer(text, return_tensors='tf')
output = model(encoded_input)
Since BertTokenizer does not support sentencepiece, AlbertTokenizer is used here.
📚 Documentation
Model Comparison
Compared with char-based models, word-based models achieve better results in most cases. Here are scores on the development set of six Chinese tasks:
Model | Score | book_review | chnsenticorp | lcqmc | tnews(CLUE) | iflytek(CLUE) | ocnli(CLUE) |
---|---|---|---|---|---|---|---|
RoBERTa-Tiny(char) | 72.3 | 83.4 | 91.4 | 81.8 | 62.0 | 55.0 | 60.3 |
RoBERTa-Tiny(word) | 74.4(+2.1) | 86.7 | 93.2 | 82.0 | 66.4 | 58.2 | 59.6 |
RoBERTa-Mini(char) | 75.9 | 85.7 | 93.7 | 86.1 | 63.9 | 58.3 | 67.4 |
RoBERTa-Mini(word) | 76.9(+1.0) | 88.5 | 94.1 | 85.4 | 66.9 | 59.2 | 67.3 |
RoBERTa-Small(char) | 76.9 | 87.5 | 93.4 | 86.5 | 65.1 | 59.4 | 69.7 |
RoBERTa-Small(word) | 78.4(+1.5) | 89.7 | 94.7 | 87.4 | 67.6 | 60.9 | 69.8 |
RoBERTa-Medium(char) | 78.0 | 88.7 | 94.8 | 88.1 | 65.6 | 59.5 | 71.2 |
RoBERTa-Medium(word) | 79.1(+1.1) | 90.0 | 95.1 | 88.0 | 67.8 | 60.6 | 73.0 |
RoBERTa-Base(char) | 79.7 | 90.1 | 95.2 | 89.2 | 67.0 | 60.9 | 75.5 |
RoBERTa-Base(word) | 80.4(+0.7) | 91.1 | 95.7 | 89.4 | 68.0 | 61.5 | 76.8 |
Hyperparameters
For each task, the best fine-tuning hyperparameters are selected from the lists below, and trained with the sequence length of 128:
- epochs: 3, 5, 8
- batch sizes: 32, 64
- learning rates: 3e-5, 1e-4, 3e-4
🔧 Technical Details
Training Data
CLUECorpusSmall is used as training data. Google's sentencepiece is used for word segmentation. The sentencepiece model is trained on CLUECorpusSmall corpus:
>>> import sentencepiece as spm
>>> spm.SentencePieceTrainer.train(input='cluecorpussmall.txt',
model_prefix='cluecorpussmall_spm',
vocab_size=100000,
max_sentence_length=1024,
max_sentencepiece_length=6,
user_defined_symbols=['[MASK]','[unused1]','[unused2]',
'[unused3]','[unused4]','[unused5]','[unused6]',
'[unused7]','[unused8]','[unused9]','[unused10]'],
pad_id=0,
pad_piece='[PAD]',
unk_id=1,
unk_piece='[UNK]',
bos_id=2,
bos_piece='[CLS]',
eos_id=3,
eos_piece='[SEP]',
train_extremely_large_corpus=True
)
Training Procedure
Models are pre-trained by UER-py on Tencent Cloud. They are pre-trained for 1,000,000 steps with a sequence length of 128 and then for 250,000 additional steps with a sequence length of 512. The same hyper-parameters are used on different model sizes.
Taking the case of word-based RoBERTa-Medium:
Stage1:
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--spm_model_path models/cluecorpussmall_spm.model \
--dataset_path cluecorpussmall_word_seq128_dataset.pt \
--processes_num 32 --seq_length 128 \
--dynamic_masking --data_processor mlm
python3 pretrain.py --dataset_path cluecorpussmall_word_seq128_dataset.pt \
--spm_model_path models/cluecorpussmall_spm.model \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_word_roberta_medium_seq128_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
--learning_rate 1e-4 --batch_size 64 \
--data_processor mlm --target mlm
Stage2:
python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
--spm_model_path models/cluecorpussmall_spm.model \
--dataset_path cluecorpussmall_word_seq512_dataset.pt \
--processes_num 32 --seq_length 512 \
--dynamic_masking --data_processor mlm
python3 pretrain.py --dataset_path cluecorpussmall_word_seq512_dataset.pt \
--spm_model_path models/cluecorpussmall_spm.model \
--pretrained_model_path models/cluecorpussmall_word_roberta_medium_seq128_model.bin-1000000 \
--config_path models/bert/medium_config.json \
--output_model_path models/cluecorpussmall_word_roberta_medium_seq512_model.bin \
--world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
--total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
--learning_rate 5e-5 --batch_size 16 \
--data_processor mlm --target mlm
Finally, the pre-trained model is converted into Huggingface's format:
python3 scripts/convert_bert_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_word_roberta_medium_seq512_model.bin-250000 \
--output_model_path pytorch_model.bin \
--layers_num 8 --type mlm
📄 License
There is no license information provided in the original README.
BibTeX entry and citation info
@article{devlin2018bert,
title={BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding},
author={Devlin, Jacob and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1810.04805},
year={2018}
}
@article{turc2019,
title={Well-Read Students Learn Better: On the Importance of Pre-training Compact Models},
author={Turc, Iulia and Chang, Ming-Wei and Lee, Kenton and Toutanova, Kristina},
journal={arXiv preprint arXiv:1908.08962v2 },
year={2019}
}
@article{zhao2019uer,
title={UER: An Open-Source Toolkit for Pre-training Models},
author={Zhao, Zhe and Chen, Hui and Zhang, Jinbin and Zhao, Xin and Liu, Tao and Lu, Wei and Chen, Xi and Deng, Haotang and Ju, Qi and Du, Xiaoyong},
journal={EMNLP-IJCNLP 2019},
pages={241},
year={2019}
}
@article{zhao2023tencentpretrain,
title={TencentPretrain: A Scalable and Flexible Toolkit for Pre-training Models of Different Modalities},
author={Zhao, Zhe and Li, Yudong and Hou, Cheng and Zhao, Jing and others},
journal={ACL 2023},
pages={217},
year={2023}

