TinyBERT-spanish-uncased-finetuned-ner Open Source Model - Suitable for Spanish Text Entity Recognition with Small Size

Tinybert Spanish Uncased Finetuned Ner

Developed by mrm8488

A named entity recognition model fine-tuned based on Spanish TinyBERT, with a size of only 55MB, suitable for entity recognition tasks in Spanish text.

Sequence Labeling Spanish#Spanish NER #Lightweight Model #Knowledge Distillation

Downloads 64

Release Time : 3/2/2022

Model Overview

This is a lightweight named entity recognition model trained using knowledge distillation technology, specifically designed to identify various named entities (such as person names, place names, organization names, etc.) in Spanish text.

Model Features

Lightweight and Efficient

The model size is only 55MB, making it more lightweight than traditional BERT models and suitable for resource-constrained environments.

Knowledge Distillation

Trained using knowledge distillation technology, significantly reducing model size while maintaining performance.

Spanish Optimization

A named entity recognition model specifically optimized for Spanish text.

Model Capabilities

Spanish Text Processing

Named Entity Recognition

Entity Classification

Use Cases

Text Analysis

News Entity Extraction

Extract key information such as person names, place names, and organization names from Spanish news.

Can identify 9 types of entity labels such as B-LOC (location) and B-PER (person name).

Social Media Analysis

Analyze mentioned entities in Spanish social media content.

🚀 Spanish TinyBERT + NER

This model is a fine - tuned version of the Spanish Tiny Bert model created using distillation for the NER downstream task on NER - C. The size of the model is only 55MB, making it a lightweight option for NER tasks.

🚀 Quick Start

✨ Features

Fine - tuned on a specific NER dataset for better performance in named entity recognition.
Significantly smaller in size compared to other models, which is beneficial for resource - constrained environments.

📦 Installation

No specific installation steps are provided in the original document.

📚 Documentation

Details of the downstream task (NER) - Dataset

Dataset: CONLL Corpora ES I preprocessed the dataset and split it into train / dev sets with an 80/20 ratio.

Dataset	# Examples
Train	8.7 K
Dev	2.2 K

Fine - tune on NER script provided by Huggingface
Labels covered:

B - LOC
B - MISC
B - ORG
B - PER
I - LOC
I - MISC
I - ORG
I - PER
O

Metrics on evaluation set

Metric	# score
F1	70.00
Precision	67.83
Recall	71.46

Comparison

Model	# F1 score	Size(MB)
bert - base - spanish - wwm - cased (BETO)	88.43	421
[bert - spanish - cased - finetuned - ner](https://huggingface.co/mrm8488/bert - spanish - cased - finetuned - ner)	90.17	420
Best Multilingual BERT	87.38	681
TinyBERT - spanish - uncased - finetuned - ner (this one)	70.00	55

💻 Usage Examples

Basic Usage

import torch
from transformers import AutoModelForTokenClassification, AutoTokenizer

id2label = {
    "0": "B-LOC",
    "1": "B-MISC",
    "2": "B-ORG",
    "3": "B-PER",
    "4": "I-LOC",
    "5": "I-MISC",
    "6": "I-ORG",
    "7": "I-PER",
    "8": "O"
}

tokenizer = AutoTokenizer.from_pretrained('mrm8488/TinyBERT-spanish-uncased-finetuned-ner')
model = AutoModelForTokenClassification.from_pretrained('mrm8488/TinyBERT-spanish-uncased-finetuned-ner')
text ="Mis amigos están pensando viajar a Londres este verano."
input_ids = torch.tensor(tokenizer.encode(text)).unsqueeze(0)

outputs = model(input_ids)
last_hidden_states = outputs[0]

for m in last_hidden_states:
  for index, n in enumerate(m):
    if(index > 0 and index <= len(text.split(" "))):
      print(text.split(" ")[index-1] + ": " + id2label[str(torch.argmax(n).item())])
      
'''
Output:
--------
Mis: O
amigos: O
están: O
pensando: O
viajar: O
a: O
Londres: B-LOC
este: O
verano.: O
'''

📄 License

No license information is provided in the original document.

Created by Manuel Romero/@mrm8488

Made with ♥ in Spain

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