File size: 2,805 Bytes
d464972
 
7cb17ce
 
d464972
7cb17ce
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
0bee447
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7cb17ce
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
---
license: apache-2.0
widget:
- text: "<|endoftext|>\nfunction getDateAfterNDay(n){\n    return moment().add(n, 'day')\n}\n// docstring\n/**"
---

## Basic info

model based [Salesforce/codegen-350M-mono](https://huggingface.co/Salesforce/codegen-350M-mono)

fine-tuned with data [codeparrot/github-code-clean](https://huggingface.co/datasets/codeparrot/github-code-clean)

data filter by JavaScript and TypeScript

## Usage

```python
from transformers import AutoTokenizer, AutoModelForCausalLM

model_type = 'kdf/javascript-docstring-generation'
tokenizer = AutoTokenizer.from_pretrained(model_type)
model = AutoModelForCausalLM.from_pretrained(model_type)

inputs = tokenizer('''<|endoftext|>
function getDateAfterNDay(n){
    return moment().add(n, 'day')
}

// docstring
/**''', return_tensors='pt')

doc_max_length = 128

generated_ids = model.generate(
    **inputs,
    max_length=inputs.input_ids.shape[1] + doc_max_length,
    do_sample=False,
    return_dict_in_generate=True,
    num_return_sequences=1,
    output_scores=True,
    pad_token_id=50256,
    eos_token_id=50256  # <|endoftext|>
)

ret = tokenizer.decode(generated_ids.sequences[0], skip_special_tokens=False)
print(ret)

```

## Prompt

You could give model a style or a specific language, for example:

```python
inputs = tokenizer('''<|endoftext|>
function add(a, b){
    return a + b;
}
// docstring
/**
  * Calculate number add.
  * @param a {number} the first number to add
  * @param b {number} the second number to add
  * @return the result of a + b
  */
<|endoftext|>
function getDateAfterNDay(n){
    return moment().add(n, 'day')
}
// docstring
/**''', return_tensors='pt')

doc_max_length = 128

generated_ids = model.generate(
    **inputs,
    max_length=inputs.input_ids.shape[1] + doc_max_length,
    do_sample=False,
    return_dict_in_generate=True,
    num_return_sequences=1,
    output_scores=True,
    pad_token_id=50256,
    eos_token_id=50256  # <|endoftext|>
)

ret = tokenizer.decode(generated_ids.sequences[0], skip_special_tokens=False)
print(ret)

inputs = tokenizer('''<|endoftext|>
function add(a, b){
    return a + b;
}
// docstring
/**
  * 计算数字相加
  * @param a {number} 第一个加数
  * @param b {number} 第二个加数
  * @return 返回 a + b 的结果
  */
<|endoftext|>
function getDateAfterNDay(n){
    return moment().add(n, 'day')
}
// docstring
/**''', return_tensors='pt')

doc_max_length = 128

generated_ids = model.generate(
    **inputs,
    max_length=inputs.input_ids.shape[1] + doc_max_length,
    do_sample=False,
    return_dict_in_generate=True,
    num_return_sequences=1,
    output_scores=True,
    pad_token_id=50256,
    eos_token_id=50256  # <|endoftext|>
)

ret = tokenizer.decode(generated_ids.sequences[0], skip_special_tokens=False)
print(ret)

```