Improve language tag
#1
by
lbourdois
- opened
README.md
CHANGED
|
@@ -1,181 +1,192 @@
|
|
| 1 |
-
|
| 2 |
-
|
| 3 |
-
|
| 4 |
-
|
| 5 |
-
|
| 6 |
-
|
| 7 |
-
|
| 8 |
-
-
|
| 9 |
-
|
| 10 |
-
-
|
| 11 |
-
|
| 12 |
-
-
|
| 13 |
-
|
| 14 |
-
-
|
| 15 |
-
|
| 16 |
-
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
| 33 |
-
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
#
|
| 43 |
-
|
| 44 |
-
|
| 45 |
-
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
|
| 59 |
-
|
| 60 |
-
|
|
| 61 |
-
|
| 62 |
-
|
| 63 |
-
|
| 64 |
-
|
| 65 |
-
|
| 66 |
-
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
|
| 70 |
-
|
| 71 |
-
|
| 72 |
-
|
| 73 |
-
|
| 74 |
-
|
| 75 |
-
|
| 76 |
-
|
| 77 |
-
|
| 78 |
-
|
| 79 |
-
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
|
| 101 |
-
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
|
| 112 |
-
|
| 113 |
-
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
|
| 120 |
-
|
| 121 |
-
|
| 122 |
-
|
| 123 |
-
|
| 124 |
-
|
| 125 |
-
|
| 126 |
-
|
| 127 |
-
|
| 128 |
-
|
| 129 |
-
|
| 130 |
-
|
| 131 |
-
|
| 132 |
-
|
| 133 |
-
|
| 134 |
-
|
| 135 |
-
|
| 136 |
-
|
| 137 |
-
|
| 138 |
-
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
-
|
| 145 |
-
|
| 146 |
-
|
| 147 |
-
|
| 148 |
-
|
| 149 |
-
|
| 150 |
-
|
| 151 |
-
|
| 152 |
-
|
| 153 |
-
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
| 165 |
-
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
base_model: Qwen/Qwen2.5-7B-Instruct
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
datasets:
|
| 6 |
+
- shibing624/chinese_text_correction
|
| 7 |
+
language:
|
| 8 |
+
- zho
|
| 9 |
+
- eng
|
| 10 |
+
- fra
|
| 11 |
+
- spa
|
| 12 |
+
- por
|
| 13 |
+
- deu
|
| 14 |
+
- ita
|
| 15 |
+
- rus
|
| 16 |
+
- jpn
|
| 17 |
+
- kor
|
| 18 |
+
- vie
|
| 19 |
+
- tha
|
| 20 |
+
- ara
|
| 21 |
+
metrics:
|
| 22 |
+
- f1
|
| 23 |
+
tags:
|
| 24 |
+
- text-generation-inference
|
| 25 |
+
widget:
|
| 26 |
+
- text: '文本纠错：
|
| 27 |
+
|
| 28 |
+
少先队员因该为老人让坐。'
|
| 29 |
+
---
|
| 30 |
+
|
| 31 |
+
[](https://hf.co/QuantFactory)
|
| 32 |
+
|
| 33 |
+
|
| 34 |
+
# QuantFactory/chinese-text-correction-7b-GGUF
|
| 35 |
+
This is quantized version of [shibing624/chinese-text-correction-7b](https://huggingface.co/shibing624/chinese-text-correction-7b) created using llama.cpp
|
| 36 |
+
|
| 37 |
+
# Original Model Card
|
| 38 |
+
|
| 39 |
+
|
| 40 |
+
|
| 41 |
+
|
| 42 |
+
# Chinese Text Correction Model
|
| 43 |
+
中文文本纠错模型chinese-text-correction-7b：用于拼写纠错、语法纠错
|
| 44 |
+
|
| 45 |
+
`shibing624/chinese-text-correction-7b` evaluate test data：
|
| 46 |
+
|
| 47 |
+
The overall performance of CSC **test**:
|
| 48 |
+
|
| 49 |
+
|input_text|predict_text|
|
| 50 |
+
|:--- |:--- |
|
| 51 |
+
|文本纠错：\n少先队员因该为老人让坐。|少先队员应该为老人让座。|
|
| 52 |
+
|
| 53 |
+
# Models
|
| 54 |
+
|
| 55 |
+
| Name | Base Model | Download |
|
| 56 |
+
|-----------------|-------------------|-----------------------------------------------------------------------|
|
| 57 |
+
| chinese-text-correction-1.5b | Qwen/Qwen2.5-1.5B-Instruct | [🤗 Hugging Face](https://huggingface.co/shibing624/chinese-text-correction-1.5b) |
|
| 58 |
+
| chinese-text-correction-1.5b-lora | Qwen/Qwen2.5-1.5B-Instruct | [🤗 Hugging Face](https://huggingface.co/shibing624/chinese-text-correction-1.5b-lora) |
|
| 59 |
+
| chinese-text-correction-7b | Qwen/Qwen2.5-7B-Instruct | [🤗 Hugging Face](https://huggingface.co/shibing624/chinese-text-correction-7b) |
|
| 60 |
+
| chinese-text-correction-7b-lora | Qwen/Qwen2.5-7B-Instruct | [🤗 Hugging Face](https://huggingface.co/shibing624/chinese-text-correction-7b-lora) |
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
### 评估结果
|
| 64 |
+
- 评估指标：F1
|
| 65 |
+
- CSC(Chinese Spelling Correction): 拼写纠错模型，表示模型可以处理音似、形似、语法等长度对齐的错误纠正
|
| 66 |
+
- CTC(CHinese Text Correction): 文本纠错模型，表示模型支持拼写、语法等长度对齐的错误纠正，还可以处理多字、少字等长度不对齐的错误纠正
|
| 67 |
+
- GPU：Tesla V100，显存 32 GB
|
| 68 |
+
|
| 69 |
+
| Model Name | Model Link | Base Model | Avg | SIGHAN-2015 | EC-LAW | MCSC | GPU/CPU | QPS |
|
| 70 |
+
|:-----------------|:------------------------------------------------------------------------------------------------------------------------|:---------------------------|:-----------|:------------|:-------|:-------|:--------|:--------|
|
| 71 |
+
| Kenlm-CSC | [shibing624/chinese-kenlm-klm](https://huggingface.co/shibing624/chinese-kenlm-klm) | kenlm | 0.3409 | 0.3147 | 0.3763 | 0.3317 | CPU | 9 |
|
| 72 |
+
| Mengzi-T5-CSC | [shibing624/mengzi-t5-base-chinese-correction](https://huggingface.co/shibing624/mengzi-t5-base-chinese-correction) | mengzi-t5-base | 0.3984 | 0.7758 | 0.3156 | 0.1039 | GPU | 214 |
|
| 73 |
+
| ERNIE-CSC | [PaddleNLP/ernie-csc](https://github.com/PaddlePaddle/PaddleNLP/tree/develop/legacy/examples/text_correction/ernie-csc) | PaddlePaddle/ernie-1.0-base-zh | 0.4353 | 0.8383 | 0.3357 | 0.1318 | GPU | 114 |
|
| 74 |
+
| MacBERT-CSC | [shibing624/macbert4csc-base-chinese](https://huggingface.co/shibing624/macbert4csc-base-chinese) | hfl/chinese-macbert-base | 0.3993 | 0.8314 | 0.1610 | 0.2055 | GPU | **224** |
|
| 75 |
+
| ChatGLM3-6B-CSC | [shibing624/chatglm3-6b-csc-chinese-lora](https://huggingface.co/shibing624/chatglm3-6b-csc-chinese-lora) | THUDM/chatglm3-6b | 0.4538 | 0.6572 | 0.4369 | 0.2672 | GPU | 3 |
|
| 76 |
+
| Qwen2.5-1.5B-CTC | [shibing624/chinese-text-correction-1.5b](https://huggingface.co/shibing624/chinese-text-correction-1.5b) | Qwen/Qwen2.5-1.5B-Instruct | 0.6802 | 0.3032 | 0.7846 | 0.9529 | GPU | 6 |
|
| 77 |
+
| Qwen2.5-7B-CTC | [shibing624/chinese-text-correction-7b](https://huggingface.co/shibing624/chinese-text-correction-7b) | Qwen/Qwen2.5-7B-Instruct | **0.8225** | 0.4917 | 0.9798 | 0.9959 | GPU | 3 |
|
| 78 |
+
|
| 79 |
+
## Usage (pycorrector)
|
| 80 |
+
|
| 81 |
+
本项目开源在`pycorrector`项目：[pycorrector](https://github.com/shibing624/pycorrector)，可支持大模型微调后用于文本纠错，通过如下命令调用：
|
| 82 |
+
|
| 83 |
+
Install package:
|
| 84 |
+
```shell
|
| 85 |
+
pip install -U pycorrector
|
| 86 |
+
```
|
| 87 |
+
|
| 88 |
+
```python
|
| 89 |
+
from pycorrector.gpt.gpt_corrector import GptCorrector
|
| 90 |
+
|
| 91 |
+
if __name__ == '__main__':
|
| 92 |
+
error_sentences = [
|
| 93 |
+
'真麻烦你了。希望你们好好的跳无',
|
| 94 |
+
'少先队员因该为老人让坐',
|
| 95 |
+
'机七学习是人工智能领遇最能体现智能的一个分知',
|
| 96 |
+
'一只小鱼船浮在平净的河面上',
|
| 97 |
+
'我的家乡是有明的渔米之乡',
|
| 98 |
+
]
|
| 99 |
+
m = GptCorrector("shibing624/chinese-text-correction-7b")
|
| 100 |
+
|
| 101 |
+
batch_res = m.correct_batch(error_sentences)
|
| 102 |
+
for i in batch_res:
|
| 103 |
+
print(i)
|
| 104 |
+
print()
|
| 105 |
+
```
|
| 106 |
+
|
| 107 |
+
## Usage (HuggingFace Transformers)
|
| 108 |
+
Without [pycorrector](https://github.com/shibing624/pycorrector), you can use the model like this:
|
| 109 |
+
|
| 110 |
+
First, you pass your input through the transformer model, then you get the generated sentence.
|
| 111 |
+
|
| 112 |
+
Install package:
|
| 113 |
+
```
|
| 114 |
+
pip install transformers
|
| 115 |
+
```
|
| 116 |
+
|
| 117 |
+
```python
|
| 118 |
+
# pip install transformers
|
| 119 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 120 |
+
checkpoint = "shibing624/chinese-text-correction-7b"
|
| 121 |
+
|
| 122 |
+
device = "cuda" # for GPU usage or "cpu" for CPU usage
|
| 123 |
+
tokenizer = AutoTokenizer.from_pretrained(checkpoint)
|
| 124 |
+
model = AutoModelForCausalLM.from_pretrained(checkpoint).to(device)
|
| 125 |
+
|
| 126 |
+
input_content = "文本纠错：\n少先队员因该为老人让坐。"
|
| 127 |
+
|
| 128 |
+
messages = [{"role": "user", "content": input_content}]
|
| 129 |
+
input_text=tokenizer.apply_chat_template(messages, tokenize=False)
|
| 130 |
+
|
| 131 |
+
print(input_text)
|
| 132 |
+
|
| 133 |
+
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
|
| 134 |
+
outputs = model.generate(inputs, max_new_tokens=1024, temperature=0, do_sample=False, repetition_penalty=1.08)
|
| 135 |
+
|
| 136 |
+
print(tokenizer.decode(outputs[0]))
|
| 137 |
+
```
|
| 138 |
+
|
| 139 |
+
output:
|
| 140 |
+
```shell
|
| 141 |
+
少先队员应该为老人让座。
|
| 142 |
+
```
|
| 143 |
+
|
| 144 |
+
|
| 145 |
+
模型文件组成：
|
| 146 |
+
```
|
| 147 |
+
shibing624/chinese-text-correction-7b
|
| 148 |
+
|-- added_tokens.json
|
| 149 |
+
|-- config.json
|
| 150 |
+
|-- generation_config.json
|
| 151 |
+
|-- merges.txt
|
| 152 |
+
|-- model.safetensors
|
| 153 |
+
|-- model.safetensors.index.json
|
| 154 |
+
|-- README.md
|
| 155 |
+
|-- special_tokens_map.json
|
| 156 |
+
|-- tokenizer_config.json
|
| 157 |
+
|-- tokenizer.json
|
| 158 |
+
`-- vocab.json
|
| 159 |
+
```
|
| 160 |
+
|
| 161 |
+
#### 训练参数：
|
| 162 |
+
|
| 163 |
+
- num_epochs: 8
|
| 164 |
+
- batch_size: 2
|
| 165 |
+
- steps: 36000
|
| 166 |
+
- eval_loss: 0.12
|
| 167 |
+
- base model: Qwen/Qwen2.5-7B-Instruct
|
| 168 |
+
- train data: [shibing624/chinese_text_correction](https://huggingface.co/datasets/shibing624/chinese_text_correction)
|
| 169 |
+
- train time: 10 days
|
| 170 |
+
- eval_loss: 
|
| 171 |
+
- train_loss: 
|
| 172 |
+
|
| 173 |
+
### 训练数据集
|
| 174 |
+
#### 中文纠错数据集
|
| 175 |
+
|
| 176 |
+
- 数据：[shibing624/chinese_text_correction](https://huggingface.co/datasets/shibing624/chinese_text_correction)
|
| 177 |
+
|
| 178 |
+
|
| 179 |
+
如果需要训练Qwen的纠错模型，请参考[https://github.com/shibing624/pycorrector](https://github.com/shibing624/pycorrector) 或者 [https://github.com/shibing624/MedicalGPT](https://github.com/shibing624/MedicalGPT)
|
| 180 |
+
|
| 181 |
+
## Citation
|
| 182 |
+
|
| 183 |
+
```latex
|
| 184 |
+
@software{pycorrector,
|
| 185 |
+
author = {Xu Ming},
|
| 186 |
+
title = {pycorrector: Implementation of language model finetune},
|
| 187 |
+
year = {2024},
|
| 188 |
+
url = {https://github.com/shibing624/pycorrector},
|
| 189 |
+
}
|
| 190 |
+
```
|
| 191 |
+
|
| 192 |
+
|