Update README.md
Browse files
README.md
CHANGED
@@ -36,7 +36,7 @@ To use the model, you need to load it with the Hugging Face Transformers library
|
|
36 |
from transformers import pipeline
|
37 |
|
38 |
# Load the model
|
39 |
-
nlp = pipeline("ner", model="
|
40 |
|
41 |
# Use the model
|
42 |
text = "ู
ุญู
ุฏ ูุงู ูู ุงูุฑุจุงุท."
|
@@ -56,6 +56,11 @@ The model is trained on a specific corpus and may not generalize well to all Mor
|
|
56 |
Performance may vary depending on text quality and tagging consistency in the dataset.
|
57 |
|
58 |
---
|
|
|
|
|
|
|
|
|
|
|
59 |
library_name: transformers
|
60 |
base_model: aubmindlab/bert-base-arabertv02
|
61 |
datasets:
|
@@ -70,56 +75,6 @@ pipeline_tag: token-classification
|
|
70 |
license: apache-2.0
|
71 |
---
|
72 |
|
73 |
-
# NER Model for Moroccan Dialect (Darija)
|
74 |
-
|
75 |
-
This model is fine-tuned for Named Entity Recognition (NER) in Moroccan Arabic (Darija). It recognizes entities such as locations, organizations, and person names in text written in Darija.
|
76 |
-
|
77 |
-
## Base Model
|
78 |
-
|
79 |
-
This model is fine-tuned from the [aubmindlab/bert-base-arabertv02](https://huggingface.co/aubmindlab/bert-base-arabertv02) model, which is optimized for Arabic NLP tasks.
|
80 |
-
|
81 |
-
## Dataset
|
82 |
-
|
83 |
-
The model is trained on the **DarNERcorp** dataset, a corpus designed for Named Entity Recognition in the Moroccan Arabic dialect.
|
84 |
-
|
85 |
-
## Task
|
86 |
-
|
87 |
-
The model is designed for the **token-classification** task, specifically Named Entity Recognition (NER).
|
88 |
-
|
89 |
-
### NER Tags
|
90 |
-
The model recognizes the following tags:
|
91 |
-
- **PER**: Person names
|
92 |
-
- **LOC**: Locations
|
93 |
-
- **ORG**: Organizations
|
94 |
-
- **MISC**: Miscellaneous entities
|
95 |
-
|
96 |
-
## Evaluation Results
|
97 |
-
|
98 |
-
The model achieves the following results on the evaluation dataset:
|
99 |
-
- **Precision**: 74.04%
|
100 |
-
- **Recall**: 85.16%
|
101 |
-
- **F1 Score**: 78.61%
|
102 |
-
|
103 |
-
## Intended Use
|
104 |
-
|
105 |
-
This model is intended for extracting named entities from Moroccan Arabic (Darija) text. It can be applied to:
|
106 |
-
- Social media content
|
107 |
-
- News articles
|
108 |
-
- Other informal or formal texts in Darija
|
109 |
-
|
110 |
-
## How to Use
|
111 |
-
|
112 |
-
You can use this model with the Hugging Face Transformers library:
|
113 |
-
|
114 |
-
```python
|
115 |
-
from transformers import pipeline
|
116 |
-
|
117 |
-
# Load the model
|
118 |
-
nlp = pipeline("ner", model="ymohannad-tazi/ner-darija-darner")
|
119 |
|
120 |
-
# Use the model
|
121 |
-
text = "ู
ุญู
ุฏ ูุงู ูู ุงูุฑุจุงุท."
|
122 |
-
result = nlp(text)
|
123 |
-
print(result)
|
124 |
|
125 |
|
|
|
36 |
from transformers import pipeline
|
37 |
|
38 |
# Load the model
|
39 |
+
nlp = pipeline("ner", model="mohannad-tazi/ner-darija-darner")
|
40 |
|
41 |
# Use the model
|
42 |
text = "ู
ุญู
ุฏ ูุงู ูู ุงูุฑุจุงุท."
|
|
|
56 |
Performance may vary depending on text quality and tagging consistency in the dataset.
|
57 |
|
58 |
---
|
59 |
+
model-index:
|
60 |
+
- name: NER_Darija_MAR_FSBM
|
61 |
+
results:
|
62 |
+
- task:
|
63 |
+
type: ner
|
64 |
library_name: transformers
|
65 |
base_model: aubmindlab/bert-base-arabertv02
|
66 |
datasets:
|
|
|
75 |
license: apache-2.0
|
76 |
---
|
77 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
78 |
|
|
|
|
|
|
|
|
|
79 |
|
80 |
|