Commit
·
9a25121
1
Parent(s):
c1deec3
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,112 @@
|
|
1 |
---
|
2 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
language:
|
3 |
+
- zh
|
4 |
+
tags:
|
5 |
+
- chatglm
|
6 |
+
- pytorch
|
7 |
+
- zh
|
8 |
+
- Text2Text-Generation
|
9 |
+
license: "apache-2.0"
|
10 |
+
widget:
|
11 |
+
- text: "介绍下北京\n答:"
|
12 |
+
|
13 |
---
|
14 |
+
|
15 |
+
# Chinese Spelling Correction LoRA Model
|
16 |
+
ChatGLM中文问答LoRA模型
|
17 |
+
|
18 |
+
`chatglm-6b-belle-zh-lora` evaluate test data:
|
19 |
+
|
20 |
+
The overall performance of chatglm-6b-belle-zh-lora on QA **test**:
|
21 |
+
|
22 |
+
|input_text|predict|
|
23 |
+
|:-- |:--- |
|
24 |
+
|介绍下北京|北京是中国的首都,是中国的政治、文化和经济中心。它位于中国的北部,拥有悠久的历史和文化遗产,如故宫、长城和天坛等。北京也是中国最著名的城市之一,有着丰富的美食、文化和娱乐活动。北京还有许多旅游景点,如颐和园、圆明园和天坛等。北京是一个充满活力和机会的城市,吸引了来自全国和世界各地的人们前来探索。|
|
25 |
+
|
26 |
+
在CSC测试集上生成结果纠错准确率高,由于是基于大模型,结果常常能带给人惊喜,不仅能纠错,还带有句子润色和改写功能。
|
27 |
+
|
28 |
+
|
29 |
+
## Usage
|
30 |
+
|
31 |
+
本项目开源在lmft项目:[lmft](https://github.com/shibing624/lmft),可支持ChatGLM模型,通过如下命令调用:
|
32 |
+
|
33 |
+
Install package:
|
34 |
+
```shell
|
35 |
+
pip install -U lmft
|
36 |
+
```
|
37 |
+
|
38 |
+
```python
|
39 |
+
from lmft import ChatGlmModel
|
40 |
+
model = ChatGlmModel("chatglm", "THUDM/chatglm-6b", lora_name="shibing624/chatglm-6b-belle-zh-lora")
|
41 |
+
r = model.predict(["介绍下北京\n答:"])
|
42 |
+
print(r) # ['北京是中国的首都,是中国的政治、文化和经济中心。...']
|
43 |
+
```
|
44 |
+
|
45 |
+
## Usage (HuggingFace Transformers)
|
46 |
+
Without [lmft](https://github.com/shibing624/lmft), you can use the model like this:
|
47 |
+
|
48 |
+
First, you pass your input through the transformer model, then you get the generated sentence.
|
49 |
+
|
50 |
+
Install package:
|
51 |
+
```
|
52 |
+
pip install transformers
|
53 |
+
```
|
54 |
+
|
55 |
+
```python
|
56 |
+
import sys
|
57 |
+
from peft import PeftModel
|
58 |
+
from transformers import AutoModel, AutoTokenizer
|
59 |
+
|
60 |
+
sys.path.append('..')
|
61 |
+
|
62 |
+
model = AutoModel.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True, device_map='auto')
|
63 |
+
model = PeftModel.from_pretrained(model, "shibing624/chatglm-6b-belle-zh-lora")
|
64 |
+
model = model.half().cuda() # fp16
|
65 |
+
tokenizer = AutoTokenizer.from_pretrained("THUDM/chatglm-6b", trust_remote_code=True)
|
66 |
+
|
67 |
+
sents = ['介绍下北京\n答:',]
|
68 |
+
for s in sents:
|
69 |
+
response = model.chat(tokenizer, s, max_length=128, eos_token_id=tokenizer.eos_token_id)
|
70 |
+
print(response)
|
71 |
+
```
|
72 |
+
|
73 |
+
output:
|
74 |
+
```shell
|
75 |
+
介绍下北京
|
76 |
+
北京是中国的首都,是中国的政治、文化和经济中心。它位于中国的北部,拥有悠久的历史和文化遗产,如故宫、长城和天坛等。北京也是中国最著名的城市之一,有着丰富的美食、文化和娱乐活动。北京还有许多旅游景点,如颐和园、圆明园和天坛等。北京是一个充满活力和机会的城市,吸引了来自全国和世界各地的人们前来探索。
|
77 |
+
```
|
78 |
+
|
79 |
+
|
80 |
+
模型文件组成:
|
81 |
+
```
|
82 |
+
chatglm-6b-belle-zh-lora
|
83 |
+
├── adapter_config.json
|
84 |
+
└── adapter_model.bin
|
85 |
+
```
|
86 |
+
|
87 |
+
|
88 |
+
### 训练数据集
|
89 |
+
|
90 |
+
1. 50万条中文ChatGPT指令Belle数据集:[BelleGroup/train_0.5M_CN](https://huggingface.co/datasets/BelleGroup/train_0.5M_CN)
|
91 |
+
2. 100万条中文ChatGPT指令Belle数据集:[BelleGroup/train_1M_CN](https://huggingface.co/datasets/BelleGroup/train_1M_CN)
|
92 |
+
3. 5万条英文ChatGPT指令Alpaca数据集:[50k English Stanford Alpaca dataset](https://github.com/tatsu-lab/stanford_alpaca#data-release)
|
93 |
+
4. 2万条中文ChatGPT指令Alpaca数据集:[shibing624/alpaca-zh](https://huggingface.co/datasets/shibing624/alpaca-zh)
|
94 |
+
5. 69万条中文指令Guanaco数据集(Belle50万条+Guanaco19万条):[Chinese-Vicuna/guanaco_belle_merge_v1.0](https://huggingface.co/datasets/Chinese-Vicuna/guanaco_belle_merge_v1.0)
|
95 |
+
|
96 |
+
|
97 |
+
如果需要训练ChatGLM模型,请参考[https://github.com/shibing624/lmft](https://github.com/shibing624/lmft)
|
98 |
+
|
99 |
+
|
100 |
+
## Citation
|
101 |
+
|
102 |
+
```latex
|
103 |
+
@software{lmft,
|
104 |
+
author = {Xu Ming},
|
105 |
+
title = {lmft: Implementation of language model finetune},
|
106 |
+
year = {2023},
|
107 |
+
url = {https://github.com/shibing624/lmft},
|
108 |
+
}
|
109 |
+
```
|
110 |
+
|
111 |
+
|
112 |
+
|