File size: 4,267 Bytes
8f3fa63 068f648 8f3fa63 |
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 |
| Feature | Description |
| --- | --- |
| **Name** | `en_spacy_ner_finetuned_news_article` |
| **Version** | `0.0.0` |
| **spaCy** | `>=3.8.3,<3.9.0` |
| **Default Pipeline** | `transformer`, `ner` |
| **Components** | `transformer`, `ner` |
| **Vectors** | 0 keys, 0 unique vectors (0 dimensions) |
| **Sources** | `https://theedgemalaysia.com/` |
| **License** | n/a |
| **Author** | [Izardy]() |
## Model Description
This is a spaCy NER model trained on English news data to identify named entities.
## Training Data
The model was trained on annotated news articles.
## Label Scheme
### Data Annotation
- spaCY training data must be labelled . The term label here referring to the entity we wish to extract from the texts. The origin of this texts could be from many sources i.e. news articles, transcripts, logs etc whereby the compilations of these texts come in various format such as parquet, json or csv. These text could contains numbers as entities depands on the user's requirement. Example of entities includes:-
> - <sup>PERSON: People, including fictional.</sup>
> - <sup>NORP: Nationalities or religious or political groups.</sup>
> - <sup>FAC: Buildings, airports, highways, bridges, etc.</sup>
> - <sup>ORG: Companies, agencies, institutions, etc.</sup>
> - <sup>GPE: Countries, cities, states.</sup>
> - <sup>LOC: Non-GPE locations, mountain ranges, bodies of water.</sup>
> - <sup>PRODUCT: Objects, vehicles, foods, etc. (Not services.)</sup>
> - <sup>EVENT: Named hurricanes, battles, wars, sports events, etc.</sup>
> - <sup>WORK_OF_ART: Titles of books, songs, etc.</sup>
> - <sup>LAW: Named documents made into laws.</sup>
> - <sup>LANGUAGE: Any named language.</sup>
> - <sup>DATE: Absolute or relative dates or periods.</sup>
> - <sup>TIME: Times smaller than a day.</sup>
> - <sup>PERCENT: Percentage, including ”%“.</sup>
> - <sup>MONEY: Monetary values, including unit.</sup>
> - <sup>QUANTITY: Measurements, as of weight or distance.</sup>
> - <sup>ORDINAL: “first”, “second”, etc.</sup>
> - <sup>CARDINAL: Numerals that do not fall under another type.</sup>
<details>
<summary>View label scheme (15 labels for 1 components)</summary>
| Component | Labels |
| --- | --- |
| **`ner`** | `EVENT`, `FAC`, `GPE`, `LAW`, `LOC`, `MONEY`, `NORP`, `ORDINAL`, `ORG`, `PERCENT`, `PERSON`, `PRODUCT`, `QUANTITY`, `TIME`, `WORK_OF_ART` |
</details>
## Result
```
=========================== Initializing pipeline ===========================
✔ Initialized pipeline
============================= Training pipeline =============================
ℹ Pipeline: ['transformer', 'ner']
ℹ Initial learn rate: 0.0
E # LOSS TRANS... LOSS NER ENTS_F ENTS_P ENTS_R SCORE
--- ------ ------------- -------- ------ ------ ------ ------
0 0 1027.60 2028.92 0.00 0.00 0.01 0.00
0 200 225652.63 77443.33 17.55 20.90 15.12 0.18
0 400 34799.92 12790.88 69.73 69.77 69.69 0.70
0 600 16755.79 9699.92 74.23 69.75 79.32 0.74
1 800 5509.15 8061.05 77.48 72.91 82.65 0.77
1 1000 5524.08 8379.61 78.38 72.42 85.41 0.78
1 1200 4490.30 7921.10 81.05 77.75 84.65 0.81
2 1400 4319.48 6579.05 81.78 79.93 83.72 0.82
2 1600 3331.55 6981.83 80.43 72.27 90.67 0.80
2 1800 2940.46 6159.03 82.06 75.51 89.86 0.82
3 2000 2940.73 5604.61 84.07 83.55 84.58 0.84
3 2200 5573.14 6000.54 85.11 82.78 87.57 0.85
3 2400 2792.99 5804.08 85.59 83.59 87.69 0.86
...
22 13600 236.77 340.90 97.73 97.00 98.47 0.98
22 13800 221.60 333.28 98.20 97.99 98.42 0.98
```
### Accuracy
| Type | Score |
| --- | --- |
| `ENTS_F` | 98.20 |
| `ENTS_P` | 97.99 |
| `ENTS_R` | 98.42 |
| `TRANSFORMER_LOSS` | 221.60 |
| `NER_LOSS` | 333.28 |
## Usage
```
import spacy
nlp = spacy.load("your-username/your-model-name")
doc = nlp("Your text here")
for ent in doc.ents:
print(ent.text, ent.label_)
```
---
--- |