Commit
·
b02fa88
1
Parent(s):
cc5e011
Update README.md
Browse files
README.md
CHANGED
|
@@ -42,10 +42,9 @@ widget:
|
|
| 42 |
|
| 43 |
|
| 44 |
* 🪄 [W&B Dashboard](https://wandb.ai/dominguesm/RestorePunctuationPTBR)
|
|
|
|
| 45 |
|
| 46 |
|
| 47 |
-
**Coming soon python package for simpler use.**
|
| 48 |
-
|
| 49 |
This is a [bert-base-portuguese-cased](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model finetuned for punctuation restoration on [WikiLingua](https://github.com/esdurmus/Wikilingua).
|
| 50 |
|
| 51 |
This model is intended for direct use as a punctuation restoration model for the general Portuguese language. Alternatively, you can use this for further fine-tuning on domain-specific texts for punctuation restoration tasks.
|
|
@@ -54,6 +53,34 @@ Model restores the following punctuations -- **[! ? . , - : ; ' ]**
|
|
| 54 |
|
| 55 |
The model also restores the upper-casing of words.
|
| 56 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 57 |
-----------------------------------------------
|
| 58 |
## 🎯 Accuracy
|
| 59 |
|
|
@@ -80,84 +107,7 @@ The model also restores the upper-casing of words.
|
|
| 80 |
| **weighted avg** | 0.96 | 0.96 | 0.96 | 1047818
|
| 81 |
|
| 82 |
-----------------------------------------------
|
| 83 |
-
## 🤷 Output
|
| 84 |
-
|
| 85 |
-
Example:
|
| 86 |
-
|
| 87 |
-
```json
|
| 88 |
-
[
|
| 89 |
-
{
|
| 90 |
-
"entity_group": "OU",
|
| 91 |
-
"score": 0.8026431202888489,
|
| 92 |
-
"word": "henrique",
|
| 93 |
-
"start": 0,
|
| 94 |
-
"end": 8
|
| 95 |
-
},
|
| 96 |
-
{
|
| 97 |
-
"entity_group": "OO",
|
| 98 |
-
"score": 0.9925149083137512,
|
| 99 |
-
"word": "foi no lago pescar com o",
|
| 100 |
-
"start": 9,
|
| 101 |
-
"end": 33
|
| 102 |
-
},
|
| 103 |
-
{
|
| 104 |
-
"entity_group": ".U",
|
| 105 |
-
"score": 0.8426014184951782,
|
| 106 |
-
"word": "pedro",
|
| 107 |
-
"start": 34,
|
| 108 |
-
"end": 39
|
| 109 |
-
},
|
| 110 |
-
{
|
| 111 |
-
"entity_group": "OU",
|
| 112 |
-
"score": 0.9519776105880737,
|
| 113 |
-
"word": "mais",
|
| 114 |
-
"start": 40,
|
| 115 |
-
"end": 44
|
| 116 |
-
},
|
| 117 |
-
{
|
| 118 |
-
"entity_group": ",O",
|
| 119 |
-
"score": 0.8551820516586304,
|
| 120 |
-
"word": "tarde",
|
| 121 |
-
"start": 45,
|
| 122 |
-
"end": 50
|
| 123 |
-
},
|
| 124 |
-
{
|
| 125 |
-
"entity_group": "OO",
|
| 126 |
-
"score": 0.9902807474136353,
|
| 127 |
-
"word": "foram para a casa do",
|
| 128 |
-
"start": 51,
|
| 129 |
-
"end": 71
|
| 130 |
-
},
|
| 131 |
-
{
|
| 132 |
-
"entity_group": "OU",
|
| 133 |
-
"score": 0.9227372407913208,
|
| 134 |
-
"word": "pedro",
|
| 135 |
-
"start": 72,
|
| 136 |
-
"end": 77
|
| 137 |
-
},
|
| 138 |
-
{
|
| 139 |
-
"entity_group": "OO",
|
| 140 |
-
"score": 0.9997054934501648,
|
| 141 |
-
"word": "fritar os",
|
| 142 |
-
"start": 78,
|
| 143 |
-
"end": 87
|
| 144 |
-
},
|
| 145 |
-
{
|
| 146 |
-
"entity_group": ".O",
|
| 147 |
-
"score": 0.9813661575317383,
|
| 148 |
-
"word": "peixes",
|
| 149 |
-
"start": 88,
|
| 150 |
-
"end": 94
|
| 151 |
-
}
|
| 152 |
-
]
|
| 153 |
-
```
|
| 154 |
-
|
| 155 |
-
This output refers to:
|
| 156 |
|
| 157 |
-
```
|
| 158 |
-
Henrique foi no lago pescar com o Pedro. Mais tarde, foram para a casa do Pedro fritar os peixes.
|
| 159 |
-
```
|
| 160 |
-
-----------------------------------------------
|
| 161 |
|
| 162 |
## 🤙 Contact
|
| 163 |
|
|
|
|
| 42 |
|
| 43 |
|
| 44 |
* 🪄 [W&B Dashboard](https://wandb.ai/dominguesm/RestorePunctuationPTBR)
|
| 45 |
+
* ⛭ [GitHub](https://github.com/DominguesM/respunct)
|
| 46 |
|
| 47 |
|
|
|
|
|
|
|
| 48 |
This is a [bert-base-portuguese-cased](https://huggingface.co/neuralmind/bert-base-portuguese-cased) model finetuned for punctuation restoration on [WikiLingua](https://github.com/esdurmus/Wikilingua).
|
| 49 |
|
| 50 |
This model is intended for direct use as a punctuation restoration model for the general Portuguese language. Alternatively, you can use this for further fine-tuning on domain-specific texts for punctuation restoration tasks.
|
|
|
|
| 53 |
|
| 54 |
The model also restores the upper-casing of words.
|
| 55 |
|
| 56 |
+
-----------------------------------------------
|
| 57 |
+
|
| 58 |
+
## 🤷 Usage
|
| 59 |
+
|
| 60 |
+
🇧🇷 easy-to-use package to restore punctuation of portuguese texts.
|
| 61 |
+
|
| 62 |
+
**Below is a quick way to use the template.**
|
| 63 |
+
|
| 64 |
+
1. First, install the package.
|
| 65 |
+
|
| 66 |
+
```
|
| 67 |
+
pip install respunct
|
| 68 |
+
```
|
| 69 |
+
|
| 70 |
+
2. Sample python code.
|
| 71 |
+
|
| 72 |
+
``` python
|
| 73 |
+
from respunct import RestorePuncts
|
| 74 |
+
|
| 75 |
+
model = RestorePuncts()
|
| 76 |
+
|
| 77 |
+
model.restore_puncts("""
|
| 78 |
+
henrique foi no lago pescar com o pedro mais tarde foram para a casa do pedro fritar os peixes""")
|
| 79 |
+
# output:
|
| 80 |
+
# Henrique foi no lago pescar com o Pedro. Mais tarde, foram para a casa do Pedro fritar os peixes.
|
| 81 |
+
|
| 82 |
+
```
|
| 83 |
+
|
| 84 |
-----------------------------------------------
|
| 85 |
## 🎯 Accuracy
|
| 86 |
|
|
|
|
| 107 |
| **weighted avg** | 0.96 | 0.96 | 0.96 | 1047818
|
| 108 |
|
| 109 |
-----------------------------------------------
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 110 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 111 |
|
| 112 |
## 🤙 Contact
|
| 113 |
|