Gpt2 Xlarge Chinese Cluecorpussmall

🚀 Chinese GPT2 Models

This set of GPT2 models is designed for Chinese text generation. It offers various model sizes to meet different needs, and can be pre - trained using different frameworks.

🚀 Quick Start

You can quickly start using the model for text generation. For example, with the GPT2 - distil model, you can use the following code:

>>> from transformers import BertTokenizer, GPT2LMHeadModel, TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("uer/gpt2-distil-chinese-cluecorpussmall")
>>> model = GPT2LMHeadModel.from_pretrained("uer/gpt2-distil-chinese-cluecorpussmall")
>>> text_generator = TextGenerationPipeline(model, tokenizer)   
>>> text_generator("这是很久之前的事情了", max_length=100, do_sample=True)
    [{'generated_text': '这是很久之前的事情了 。 我 现 在 想 起 来 就 让 自 己 很 伤 心 ， 很 失 望 。 我 现 在 想 到 ， 我 觉 得 大 多 数 人 的 生 活 比 我 的 生 命 还 要 重 要 ， 对 一 些 事 情 的 看 法 ， 对 一 些 人 的 看 法 ， 都 是 在 发 泄 。 但 是 ， 我 们 的 生 活 是 需 要 一 个 信 用 体 系 的 。 我 不 知'}]

✨ Features

• Multiple Model Sizes: There are five different sizes of GPT2 models available, including GPT2 - distil, GPT2, GPT2 - medium, GPT2 - large, and GPT2 - xlarge, which can be selected according to specific requirements.
• Diverse Pre - training Frameworks: Except for the GPT2 - xlarge model, other models are pre - trained by [UER - py](https://github.com/dbiir/UER - py/). The GPT2 - xlarge model is pre - trained by TencentPretrain, which supports models with parameters above one billion and extends to a multimodal pre - training framework.

📦 Installation

You can download the set of Chinese GPT2 models either from the [UER - py Modelzoo page](https://github.com/dbiir/UER - py/wiki/Modelzoo), or via HuggingFace from the links below:

💻 Usage Examples

Basic Usage

>>> from transformers import BertTokenizer, GPT2LMHeadModel, TextGenerationPipeline
>>> tokenizer = BertTokenizer.from_pretrained("uer/gpt2-distil-chinese-cluecorpussmall")
>>> model = GPT2LMHeadModel.from_pretrained("uer/gpt2-distil-chinese-cluecorpussmall")
>>> text_generator = TextGenerationPipeline(model, tokenizer)   
>>> text_generator("这是很久之前的事情了", max_length=100, do_sample=True)
    [{'generated_text': '这是很久之前的事情了 。 我 现 在 想 起 来 就 让 自 己 很 伤 心 ， 很 失 望 。 我 现 在 想 到 ， 我 觉 得 大 多 数 人 的 生 活 比 我 的 生 命 还 要 重 要 ， 对 一 些 事 情 的 看 法 ， 对 一 些 人 的 看 法 ， 都 是 在 发 泄 。 但 是 ， 我 们 的 生 活 是 需 要 一 个 信 用 体 系 的 。 我 不 知'}]

📚 Documentation

Model description

The set of GPT2 models, except for GPT2 - xlarge model, are pre - trained by [UER - py](https://github.com/dbiir/UER - py/), which is introduced in this paper. The GPT2 - xlarge model is 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. Besides, the other models could also be pre - trained by TencentPretrain.

Training data

CLUECorpusSmall is used as training data.

Training procedure

For models pre - trained by UER - py (take GPT2 - distil as an example)

Stage1:

python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
                      --vocab_path models/google_zh_vocab.txt \
                      --dataset_path cluecorpussmall_lm_seq128_dataset.pt \
                      --seq_length 128 --processes_num 32 --data_processor lm 

python3 pretrain.py --dataset_path cluecorpussmall_lm_seq128_dataset.pt \
                    --vocab_path models/google_zh_vocab.txt \
                    --config_path models/gpt2/distil_config.json \
                    --output_model_path models/cluecorpussmall_gpt2_distil_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

Stage2:

python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
                      --vocab_path models/google_zh_vocab.txt \
                      --dataset_path cluecorpussmall_lm_seq1024_dataset.pt \
                      --seq_length 1024 --processes_num 32 --data_processor lm 

python3 pretrain.py --dataset_path cluecorpussmall_lm_seq1024_dataset.pt \
                    --vocab_path models/google_zh_vocab.txt \
                    --pretrained_model_path models/cluecorpussmall_gpt2_distil_seq128_model.bin-1000000 \
                    --config_path models/gpt2/distil_config.json \
                    --output_model_path models/cluecorpussmall_gpt2_distil_seq1024_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

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

python3 scripts/convert_gpt2_from_uer_to_huggingface.py --input_model_path models/cluecorpussmall_gpt2_distil_seq1024_model.bin-250000 \
                                                        --output_model_path pytorch_model.bin \
                                                        --layers_num 6

For GPT2 - xlarge model (using TencetPretrain)

Stage1:

python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
                      --vocab_path models/google_zh_vocab.txt \
                      --dataset_path cluecorpussmall_lm_seq128_dataset.pt \
                      --seq_length 128 --processes_num 32 --data_processor lm

deepspeed pretrain.py --deepspeed --deepspeed_config models/deepspeed_config.json \
                      --dataset_path corpora/cluecorpussmall_lm_seq128_dataset.pt \
                      --vocab_path models/google_zh_vocab.txt \
                      --config_path models/gpt2/xlarge_config.json \
                      --output_model_path models/cluecorpussmall_gpt2_xlarge_seq128_model \
                      --world_size 8 --batch_size 64 \
                      --total_steps 1000000 --save_checkpoint_steps 100000 --report_steps 50000 \
                      --deepspeed_checkpoint_activations --deepspeed_checkpoint_layers_num 24

Before stage2, extract fp32 consolidated weights from a zero 2 and 3 DeepSpeed checkpoints:

python3 models/cluecorpussmall_gpt2_xlarge_seq128_model/zero_to_fp32.py models/cluecorpussmall_gpt2_xlarge_seq128_model/ \
                                                                        models/cluecorpussmall_gpt2_xlarge_seq128_model.bin

Stage2:

python3 preprocess.py --corpus_path corpora/cluecorpussmall.txt \
                      --vocab_path models/google_zh_vocab.txt \
                      --dataset_path cluecorpussmall_lm_seq1024_dataset.pt \
                      --seq_length 1024 --processes_num 32 --data_processor lm

deepspeed pretrain.py --deepspeed --deepspeed_config models/deepspeed_config.json \
                      --dataset_path corpora/cluecorpussmall_lm_seq1024_dataset.pt \
                      --vocab_path models/google_zh_vocab.txt \
                      --config_path models/gpt2/xlarge_config.json \
                      --pretrained_model_path models/cluecorpussmall_gpt2_xlarge_seq128_model.bin \
                      --output_model_path models/cluecorpussmall_gpt2_xlarge_seq1024_model \
                      --world_size 8 --batch_size 16 --learning_rate 5e-5 \
                      --total_steps 250000 --save_checkpoint_steps 50000 --report_steps 10000 \
                      --deepspeed_checkpoint_activations --deepspeed_checkpoint_layers_num 6

Then, extract fp32 consolidated weights from a zero 2 and 3 DeepSpeed checkpoints:

python3 models/cluecorpussmall_gpt2_xlarge_seq1024_model/zero_to_fp32.py models/cluecorpussmall_gpt2_xlarge_seq1024_model/ \
                                                                         models/cluecorpussmall_gpt2_xlarge_seq1024_model.bin

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

python3 scripts/convert_gpt2_from_tencentpretrain_to_huggingface.py --input_model_path models/cluecorpussmall_gpt2_xlarge_seq1024_model.bin \
                                                                    --output_model_path pytorch_model.bin \
                                                                    --layers_num 48

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}