roberta-fa-zwnj-base-ner Open-Source Model - Accurately Identify 10 Persian Entity Types

Roberta Fa Zwnj Base Ner

Developed by HooshvareLab

This model is a RoBERTa model fine-tuned for Persian Named Entity Recognition (NER) tasks, supporting the identification of 10 entity types.

Sequence Labeling Other#Persian Named Entity Recognition #Multi-entity Type Recognition #Mixed Dataset Training

Downloads 102

Release Time : 3/2/2022

Model Overview

A Persian Named Entity Recognition model based on the RoBERTa architecture, trained on mixed NER datasets, capable of identifying 10 types of entities including dates, events, facilities, locations, etc.

Model Features

Multi-type Entity Recognition

Supports the identification of 10 different types of named entities, including dates, events, facilities, etc.

Mixed Dataset Training

Trained using multiple Persian NER datasets such as ARMAN, PEYMA, and WikiANN.

High Accuracy

Achieves an overall F1 score of 0.955 on the test set, with most entity types exceeding an F1 score of 0.95.

Model Capabilities

Persian Text Processing

Named Entity Recognition

Entity Classification

Use Cases

Information Extraction

News Entity Extraction

Extracts key information such as person names, organization names, and locations from Persian news texts.

Accurately identifies 2,646 person name entities with an F1 score of 0.958

Financial Document Analysis

Identifies key numerical information such as currencies and percentages in financial documents.

Currency recognition F1 score of 0.928, percentage recognition F1 score of 0.984

Knowledge Graph Construction

Entity Relation Extraction

Serves as a preliminary step for knowledge graph construction by identifying various entities in texts.

🚀 RobertaNER

This model is fine - tuned for the Named Entity Recognition (NER) task. It uses a mixed NER dataset collected from ARMAN, PEYMA, and WikiANN, covering ten types of entities.

✨ Features

This model can recognize ten types of entities:

Date (DAT)
Event (EVE)
Facility (FAC)
Location (LOC)
Money (MON)
Organization (ORG)
Percent (PCT)
Person (PER)
Product (PRO)
Time (TIM)

📦 Installation

Installing requirements

pip install transformers

💻 Usage Examples

Basic Usage

from transformers import AutoTokenizer
from transformers import AutoModelForTokenClassification  # for pytorch
from transformers import TFAutoModelForTokenClassification  # for tensorflow
from transformers import pipeline


model_name_or_path = "HooshvareLab/roberta-fa-zwnj-base-ner" 
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path)
model = AutoModelForTokenClassification.from_pretrained(model_name_or_path)  # Pytorch
# model = TFAutoModelForTokenClassification.from_pretrained(model_name_or_path)  # Tensorflow

nlp = pipeline("ner", model=model, tokenizer=tokenizer)
example = "در سال ۲۰۱۳ درگذشت و آندرتیکر و کین برای او مراسم یادبود گرفتند."

ner_results = nlp(example)
print(ner_results)

📚 Documentation

Dataset Information

Property	Details
Model Type	Roberta fine - tuned for NER
Training Data	Mixed NER dataset from ARMAN, PEYMA, and WikiANN

	Records	B-DAT	B-EVE	B-FAC	B-LOC	B-MON	B-ORG	B-PCT	B-PER	B-PRO	B-TIM	I-DAT	I-EVE	I-FAC	I-LOC	I-MON	I-ORG	I-PCT	I-PER	I-PRO	I-TIM
Train	29133	1423	1487	1400	13919	417	15926	355	12347	1855	150	1947	5018	2421	4118	1059	19579	573	7699	1914	332
Valid	5142	267	253	250	2362	100	2651	64	2173	317	19	373	799	387	717	270	3260	101	1382	303	35
Test	6049	407	256	248	2886	98	3216	94	2646	318	43	568	888	408	858	263	3967	141	1707	296	78

Evaluation

The following tables summarize the scores obtained by the model overall and per each class.

Overall

Model	accuracy	precision	recall	f1
Roberta	0.994849	0.949816	0.960235	0.954997

Per entities

	number	precision	recall	f1
DAT	407	0.844869	0.869779	0.857143
EVE	256	0.948148	1.000000	0.973384
FAC	248	0.957529	1.000000	0.978304
LOC	2884	0.965422	0.968100	0.966759
MON	98	0.937500	0.918367	0.927835
ORG	3216	0.943662	0.958333	0.950941
PCT	94	1.000000	0.968085	0.983784
PER	2646	0.957030	0.959562	0.958294
PRO	318	0.963636	1.000000	0.981481
TIM	43	0.739130	0.790698	0.764045

❓ Questions?

Post a Github issue on the ParsNER Issues repo.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご