tomaarsen
/

span-marker-xlm-roberta-large-conll03

Token Classification

named-entity-recognition

Model card Files Files and versions Community

span-marker-xlm-roberta-large-conll03 / train.py

tomaarsen's picture

tomaarsen HF staff

Upload train.py with huggingface_hub

f489e24 11 months ago

raw history blame contribute delete

No virus

1.88 kB

	from datasets import load_dataset
	from span_marker import SpanMarkerModel, Trainer
	from transformers import TrainingArguments


	def main() -> None:
	# Load the dataset, ensure "tokens" and "ner_tags" columns, and get a list of labels
	dataset = load_dataset("conll2003")
	labels = dataset["train"].features["ner_tags"].feature.names

	# Initialize a SpanMarker model using a pretrained BERT-style encoder
	model_name = "xlm-roberta-large"
	model = SpanMarkerModel.from_pretrained(
	model_name,
	labels=labels,
	# SpanMarker hyperparameters:
	model_max_length=128,
	marker_max_length=64,
	entity_max_length=6,
	)

	# Prepare the 🤗 transformers training arguments
	args = TrainingArguments(
	output_dir="models/span_marker_xlm_roberta_large_conll03",
	# Training Hyperparameters:
	learning_rate=1e-5,
	per_device_train_batch_size=4,
	per_device_eval_batch_size=4,
	gradient_accumulation_steps=2,
	num_train_epochs=3,
	weight_decay=0.01,
	warmup_ratio=0.1,
	bf16=True,
	# Other Training parameters
	logging_first_step=True,
	logging_steps=50,
	evaluation_strategy="steps",
	save_strategy="steps",
	eval_steps=1000,
	dataloader_num_workers=2,
	)

	# Initialize the trainer using our model, training args & dataset, and train
	trainer = Trainer(
	model=model,
	args=args,
	train_dataset=dataset["train"],
	eval_dataset=dataset["validation"],
	)
	trainer.train()
	trainer.save_model("models/span_marker_xlm_roberta_large_conll03/checkpoint-final")

	# Compute & save the metrics on the test set
	metrics = trainer.evaluate(dataset["test"], metric_key_prefix="test")
	trainer.save_metrics("test", metrics)


	if __name__ == "__main__":
	main()