Jerado/span-marker-roberta-large-enron

SpanMarker with roberta-large on Jerado/enron_intangibles_ner

This is a SpanMarker model trained on the Jerado/enron_intangibles_ner dataset that can be used for Named Entity Recognition. This SpanMarker model uses roberta-large as the underlying encoder.

Model Details

Model Description

• Model Type: SpanMarker
• Encoder: roberta-large
• Maximum Sequence Length: 256 tokens
• Maximum Entity Length: 6 words
• Training Dataset: Jerado/enron_intangibles_ner
• Language: en
• License: apache-2.0

Model Sources

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

Model Labels

Label	Examples
Intangible	"deal", "sample EES deal", "Enpower system"

Evaluation

Metrics

Label	Precision	Recall	F1
all	0.4286	0.45	0.4390
Intangible	0.4286	0.45	0.4390

Uses

Direct Use for Inference

from span_marker import SpanMarkerModel

# Download from the 🤗 Hub
model = SpanMarkerModel.from_pretrained("span_marker_model_id")
# Run inference
entities = model.predict("It seems that there is a single significant policy concern for the ASIC policy committee.")

Downstream Use

You can finetune this model on your own dataset.

Click to expand

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	1	19.8706	216
Entities per sentence	0	0.1865	6

Training Hyperparameters

• learning_rate: 1e-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: 11
• mixed_precision_training: Native AMP

Training Results

Epoch	Step	Validation Loss	Validation Precision	Validation Recall	Validation F1	Validation Accuracy
3.3557	500	0.0075	0.4444	0.1667	0.2424	0.9753
6.7114	1000	0.0084	0.5714	0.3333	0.4211	0.9793
10.0671	1500	0.0098	0.6111	0.4583	0.5238	0.9815

Framework Versions

• Python: 3.10.12
• SpanMarker: 1.5.0
• Transformers: 4.40.0
• PyTorch: 2.2.1+cu121
• Datasets: 2.19.0
• Tokenizers: 0.19.1

Citation

BibTeX

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