Hugging Face's logo

p208p2002
/
bloom-1b1-zh-error-correction-dpo

Text Generation
Transformers
Safetensors
Chinese
bloom
text-generation-inference
Inference Endpoints
Model card Files Community
bloom-1b1-zh-error-correction-dpo
File size: 5,381 Bytes 
756fdd3
 
 
 
 
 
6732c6e
bf7144f
 
6732c6e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
 
---
language:
- zh
datasets:
- p208p2002/zhtw-sentence-error-correction
---
# DPO Chinese Error Correction Model

使用DPO訓練的中文糾錯模型。

### Usage
```python
from transformers import AutoTokenizer, AutoModelForCausalLM, LlamaForCausalLM,AddedToken
import sys

mode_id = "p208p2002/bloom-1b1-zh-error-correction-dpo"
model: LlamaForCausalLM = AutoModelForCausalLM.from_pretrained("p208p2002/bloom-1b1-zh-error-correction-dpo")
tokenizer = AutoTokenizer.from_pretrained("p208p2002/bloom-1b1-zh-error-correction-dpo")

test_texts = [
    "為了潔約能源請隨守關閉沒有使用的電器",
    "今天新情很好",
    "你快樂我也很高心",
    "但不再算再找實習生了",
    "今天太陽很大要注意篩傷",
    "你要不要和我依起去台北",
    "清晨六點終太陽會升起",
    "傾城六點鐘太陽會升起",
    "鍋馬路時你應該要注意虹綠燈",
    "他正在學學彈吉他",
    "下樓梯請注意階梯",
    "此信件為系統自動發送之通知",
    "此信件為系統自動發送知通知",
    "如為誤傳也請立即刪除本郵件並通知寄件者"
]
for text in test_texts:
    inputs = tokenizer(
        f"{tokenizer.bos_token}{text} {tokenizer.eos_token}\n {tokenizer.bos_token}",
        return_tensors="pt",
        add_special_tokens=False
    )["input_ids"]

    out = model.generate(
        inputs,
        max_new_tokens=20,
    )
    decode_out = tokenizer.decode(out[0])

    input_text,output_text = decode_out.split("\n") 
    input_text = input_text.strip()
    output_text = output_text.strip()

    print("input :",input_text)
    print("output:",output_text)
    print('-'*30)
```
```
input: <s>為了潔約能源請隨守關閉沒有使用的電器 </s>
output: <s>為了節約能源請隨時關閉沒有使用的電器 </s>
------------------------------
input: <s>今天新情很好 </s>
output: <s>今天心情很好 </s>
------------------------------
input: <s>你快樂我也很高心 </s>
output: <s>你快樂我也很高興 </s>
------------------------------
input: <s>但不再算再找實習生了 </s>
output: <s>但不再去找實習生了 </s>
------------------------------
input: <s>今天太陽很大要注意篩傷 </s>
output: <s>今天太陽很大要注意一下 </s>
------------------------------
input: <s>你要不要和我依起去台北 </s>
output: <s>你要不要和我一起去台北 </s>
------------------------------
input: <s>清晨六點終太陽會升起 </s>
output: <s>清晨六點鐘太陽會升起 </s>
------------------------------
input: <s>傾城六點鐘太陽會升起 </s>
output: <s>凌晨六點鐘太陽會升起 </s>
------------------------------
input: <s>鍋馬路時你應該要注意虹綠燈 </s>
output: <s>過馬路時你應該要注意紅綠燈 </s>
------------------------------
input: <s>他正在學學彈吉他 </s>
output: <s>他正在學習彈吉他 </s>
------------------------------
input: <s>下樓梯請注意階梯 </s>
output: <s>下樓梯請注意階梯 </s>
------------------------------
input: <s>此信件為系統自動發送之通知 </s>
output: <s>此信件為系統自動發送之通知 </s>
------------------------------
input: <s>此信件為系統自動發送知通知 </s>
output: <s>此信件為系統自動發送通知 </s>
------------------------------
input: <s>如為誤傳也請立即刪除本郵件並通知寄件者 </s>
output: <s>如為誤傳也請立即刪除本郵件並通知寄件者 </s>
------------------------------
input : <s>為了潔約能源請隨守關閉沒有使用的電器 </s>
output: <s>為了節約能源請隨時關閉沒有使用的電器 </s>
------------------------------
input : <s>今天新情很好 </s>
output: <s>今天心情很好 </s>
------------------------------
input : <s>你快樂我也很高心 </s>
output: <s>你快樂我也很高興 </s>
------------------------------
input : <s>但不再算再找實習生了 </s>
output: <s>但不再去找實習生了 </s>
------------------------------
input : <s>今天太陽很大要注意篩傷 </s>
output: <s>今天太陽很大要注意一下 </s>
------------------------------
input : <s>你要不要和我依起去台北 </s>
output: <s>你要不要和我一起去台北 </s>
------------------------------
input : <s>清晨六點終太陽會升起 </s>
output: <s>清晨六點鐘太陽會升起 </s>
------------------------------
input : <s>傾城六點鐘太陽會升起 </s>
output: <s>凌晨六點鐘太陽會升起 </s>
------------------------------
input : <s>鍋馬路時你應該要注意虹綠燈 </s>
output: <s>過馬路時你應該要注意紅綠燈 </s>
------------------------------
input : <s>他正在學學彈吉他 </s>
output: <s>他正在學習彈吉他 </s>
------------------------------
input : <s>下樓梯請注意階梯 </s>
output: <s>下樓梯請注意階梯 </s>
------------------------------
input : <s>此信件為系統自動發送之通知 </s>
output: <s>此信件為系統自動發送之通知 </s>
------------------------------
input : <s>此信件為系統自動發送知通知 </s>
output: <s>此信件為系統自動發送通知 </s>
------------------------------
input : <s>如為誤傳也請立即刪除本郵件並通知寄件者 </s>
output: <s>如為誤傳也請立即刪除本郵件並通知寄件者 </s>
------------------------------
```