This is a SpanMarker model trained on the conll2002 dataset that can be used for Named Entity Recognition. This SpanMarker model uses bert-base-cased as the underlying encoder.

Model Details

Model Description

Model Type: SpanMarker
Encoder: bert-base-cased
Maximum Sequence Length: 256 tokens
Maximum Entity Length: 8 words
Training Dataset: conll2002
Language: es
License: cc-by-sa-4.0

Model Sources

Repository: SpanMarker on GitHub
Thesis: SpanMarker For Named Entity Recognition

Model Labels

Label	Examples
LOC	"Victoria", "Australia", "Melbourne"
MISC	"Ley", "Ciudad", "CrimeNet"
ORG	"Tribunal Supremo", "EFE", "Commonwealth"
PER	"Abogado General del Estado", "Daryl Williams", "Abogado General"

Evaluation

Metrics

Label	Precision	Recall	F1
all	0.8331	0.8074	0.8201
LOC	0.8471	0.7759	0.8099
MISC	0.7092	0.4264	0.5326
ORG	0.7854	0.8558	0.8191
PER	0.9471	0.9329	0.9400

Uses

Direct Use for Inference

python

from span_marker import SpanMarkerModel

# Download from the 🤗 Hub
model = SpanMarkerModel.from_pretrained("span_marker_model_id")
# Run inference
entities = model.predict("(SV2147) PP: PROBLEMAS INTERNOS PSOE INTERFIEREN EN POLITICA DE LA JUNTA Córdoba (EFE).")

Downstream Use

You can finetune this model on your own dataset.

Click to expand

python

from span_marker import SpanMarkerModel, Trainer

# Download from the 🤗 Hub
model = SpanMarkerModel.from_pretrained("span_marker_model_id")

# Specify a Dataset with "tokens" and "ner_tag" columns
dataset = load_dataset("conll2003") # For example CoNLL2003

# Initialize a Trainer using the pretrained model & dataset
trainer = Trainer(
    model=model,
    train_dataset=dataset["train"],
    eval_dataset=dataset["validation"],
)
trainer.train()
trainer.save_model("span_marker_model_id-finetuned")

Training Details

Training Set Metrics

Training set	Min	Median	Max
Sentence length	0	31.8014	1238
Entities per sentence	0	2.2583	160

Training Hyperparameters

learning_rate: 5e-05
train_batch_size: 4
eval_batch_size: 4
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 1
mixed_precision_training: Native AMP

Training Results

Epoch	Step	Validation Loss	Validation Precision	Validation Recall	Validation F1	Validation Accuracy
0.1164	200	0.0260	0.6907	0.5358	0.6035	0.9264
0.2328	400	0.0199	0.7567	0.6384	0.6925	0.9414
0.3491	600	0.0176	0.7773	0.7273	0.7515	0.9563
0.4655	800	0.0157	0.8066	0.7598	0.7825	0.9601
0.5819	1000	0.0158	0.8031	0.7413	0.7710	0.9605
0.6983	1200	0.0156	0.7975	0.7598	0.7782	0.9609
0.8147	1400	0.0139	0.8210	0.7615	0.7901	0.9625
0.9310	1600	0.0129	0.8426	0.7848	0.8127	0.9651

Framework Versions

Python: 3.10.12
SpanMarker: 1.5.0
Transformers: 4.38.2
PyTorch: 2.2.1+cu121
Datasets: 2.18.0
Tokenizers: 0.15.2

Citation

BibTeX

@software{Aarsen_SpanMarker,
    author = {Aarsen, Tom},
    license = {Apache-2.0},
    title = {{SpanMarker for Named Entity Recognition}},
    url = {https://github.com/tomaarsen/SpanMarkerNER}
}

sepulm01/span-marker-bert-base-conll2002-es

SpanMarker with bert-base-cased on conll2002