Realise For Csc

🚀 ReaLiSe-for-csc

A Chinese Spell Checking (CSC) model

This model is derived from the one provided in the ReaLiSe source code.

The original paper can be found at: https://arxiv.org/abs/2105.12306

The official code of the original paper is available at: https://github.com/DaDaMrX/ReaLiSe

The performance of this model on SIGHAN2015 is as follows:

🚀 Quick Start

📦 Installation

Install the dependencies:

!pip install transformers
!pip install pypinyin
!pip install boto3

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("iioSnail/ReaLiSe-for-csc", trust_remote_code=True)
model = AutoModel.from_pretrained("iioSnail/ReaLiSe-for-csc", trust_remote_code=True)

inputs = tokenizer(["我是炼习时长两念半的个人练习生蔡徐坤"], return_tensors='pt')
output_hidden = model(**inputs).logits
print(''.join(tokenizer.convert_ids_to_tokens(output_hidden.argmax(-1)[0, 1:-1])))

Output:

我是练习时长两年半的个人练习生蔡徐坤

Advanced Usage

You can also use the predict method encapsulated in this model.

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("iioSnail/ReaLiSe-for-csc", trust_remote_code=True)
model = AutoModel.from_pretrained("iioSnail/ReaLiSe-for-csc", trust_remote_code=True)

model.set_tokenizer(tokenizer)  # Call this method before using the predict method
print(model.predict("我是练习时长两念半的鸽仁练习生蔡徐坤"))
print(model.predict(["我是练习时长两念半的鸽仁练习生蔡徐坤", "喜换唱跳、rap 和 蓝球"]))

Output:

我是练习时长两年半的各仁练习生蔡徐坤
['我是练习时长两年半的各仁练习生蔡徐坤', '喜欢唱跳、rap 和 蓝球']

📚 Documentation

Model Training

from transformers import AutoTokenizer, AutoModel

tokenizer = AutoTokenizer.from_pretrained("iioSnail/ReaLiSe-for-csc", trust_remote_code=True)
model = AutoModel.from_pretrained("iioSnail/ReaLiSe-for-csc", trust_remote_code=True)

inputs = tokenizer(["我是炼习时长两念半的个人练习生蔡徐坤", "喜换唱跳rap蓝球"],
                   text_target=["我是练习时长两年半的个人练习生蔡徐坤", "喜欢唱跳rap篮球"],
                   padding=True,
                   return_tensors='pt')
loss = model(**inputs).loss
print("loss:", loss)
loss.backward()

Output:

loss: tensor(0.6515, grad_fn=<NllLossBackward0>)

Frequently Asked Questions

1. Network issues, e.g., Connection Error

Solution: Download the model to your local machine for use. You can refer to this blog for the batch download solution.

2. An error occurs when using the model downloaded locally: ModuleNotFoundError: No module named 'transformers_modules.iioSnail/ReaLiSe-for-csc'

Solution: Change iioSnail/ChineseBERT-for-csc to iioSnail\ChineseBERT-for-csc, or upgrade transformers.

📄 License

This project is licensed under the AFL-3.0 license.