Gpt2 Chinese Couplet

🚀 Chinese Couplet GPT2 Model

This is a model for generating Chinese couplets, pre - trained using advanced pre - training frameworks.

🚀 Quick Start

You can quickly start using this model to generate Chinese couplets. The following will introduce the model's description, usage, training data, and training process.

✨ Features

• Powerful Pre - training Frameworks: The model can be pre - trained by [UER - py](https://github.com/dbiir/UER - py/) and TencentPretrain, which are introduced in this paper and this paper respectively.
• Convenient Download Channels: You can download the model from multiple sources, such as the [UER - py Modelzoo page](https://github.com/dbiir/UER - py/wiki/Modelzoo), [GPT2 - Chinese Github page](https://github.com/Morizeyao/GPT2 - Chinese), or via HuggingFace from the link [gpt2 - chinese - couplet](https://huggingface.co/uer/gpt2 - chinese - couplet).

📦 Installation

There is no specific installation process described in the original text. You can directly download the model from the provided sources.

💻 Usage Examples

Basic Usage

You can use the model directly with a pipeline for text generation. When the parameter skip_special_tokens is True:

>>> from transformers import BertTokenizer, GPT2LMHeadModel, TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("uer/gpt2-chinese-couplet")
>>> model = GPT2LMHeadModel.from_pretrained("uer/gpt2-chinese-couplet")
>>> text_generator = TextGenerationPipeline(model, tokenizer)   
>>> text_generator("[CLS]丹 枫 江 冷 人 初 去 -", max_length=25, do_sample=True)
    [{'generated_text': '[CLS]丹 枫 江 冷 人 初 去 - 黄 叶 声 从 天 外 来 阅 旗'}]

Advanced Usage

When the parameter skip_special_tokens is False:

>>> from transformers import BertTokenizer, GPT2LMHeadModel, TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("uer/gpt2-chinese-couplet")
>>> model = GPT2LMHeadModel.from_pretrained("uer/gpt2-chinese-couplet")
>>> text_generator = TextGenerationPipeline(model, tokenizer)   
>>> text_generator("[CLS]丹 枫 江 冷 人 初 去 -", max_length=25, do_sample=True)
    [{'generated_text': '[CLS]丹 枫 江 冷 人 初 去 - 黄 叶 声 我 酒 不 辞 [SEP] [SEP] [SEP] [SEP] [SEP] [SEP] [SEP] [SEP] [SEP]'}]

📚 Documentation

Model Description

The model is pre - trained by [UER - py](https://github.com/dbiir/UER - py/), which is introduced in this paper. Besides, the model could also be pre - trained by TencentPretrain introduced in this paper, which inherits UER - py to support models with parameters above one billion, and extends it to a multimodal pre - training framework.

The model is used to generate Chinese couplets. You can download the model from the [UER - py Modelzoo page](https://github.com/dbiir/UER - py/wiki/Modelzoo), or [GPT2 - Chinese Github page](https://github.com/Morizeyao/GPT2 - Chinese), or via HuggingFace from the link [gpt2 - chinese - couplet](https://huggingface.co/uer/gpt2 - chinese - couplet).

Since the parameter skip_special_tokens is used in the pipelines.py, special tokens such as [SEP], [UNK] will be deleted, the output results of Hosted inference API (right) may not be properly displayed.

Training Data

Training data contains 700,000 Chinese couplets which are collected by [couplet - clean - dataset](https://github.com/v - zich/couplet - clean - dataset).

Training Procedure

The model is pre - trained by [UER - py](https://github.com/dbiir/UER - py/) on Tencent Cloud. We pre - train 25,000 steps with a sequence length of 64.

python3 preprocess.py --corpus_path corpora/couplet.txt \
                      --vocab_path models/google_zh_vocab.txt \
                      --dataset_path couplet_dataset.pt --processes_num 16 \
                      --seq_length 64 --data_processor lm 

python3 pretrain.py --dataset_path couplet_dataset.pt \
                    --vocab_path models/google_zh_vocab.txt \
                    --config_path models/gpt2/config.json \
                    --output_model_path models/couplet_gpt2_model.bin \
                    --world_size 8 --gpu_ranks 0 1 2 3 4 5 6 7 \
                    --total_steps 25000 --save_checkpoint_steps 5000 --report_steps 1000 \
                    --learning_rate 5e-4 --batch_size 64

Finally, we convert the pre - trained model into Huggingface's format:

python3 scripts/convert_gpt2_from_uer_to_huggingface.py --input_model_path models/couplet_gpt2_model.bin-25000 \
                                                        --output_model_path pytorch_model.bin \
                                                        --layers_num 12

BibTeX entry and citation info

@article{radford2019language,
  title={Language Models are Unsupervised Multitask Learners},
  author={Radford, Alec and Wu, Jeff and Child, Rewon and Luan, David and Amodei, Dario and Sutskever, Ilya},
  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}
}

📄 License

The original text does not provide license information, so this section is skipped.

🔧 Technical Details

The original text does not provide in - depth technical details, so this section is skipped.

⚠️ Important Note

Since the parameter skip_special_tokens is used in the pipelines.py, special tokens such as [SEP], [UNK] will be deleted, the output results of Hosted inference API (right) may not be properly displayed.