tomaarsen's picture
tomaarsen HF staff
Upload train.py with huggingface_hub
2bc8e1c
from datasets import load_dataset
from transformers import TrainingArguments
from span_marker import SpanMarkerModel, Trainer
def main() -> None:
# Load the dataset, ensure "tokens" and "ner_tags" columns, and get a list of labels
train_dataset = load_dataset("P3ps/Cross_ner", split="train")
test_dataset = load_dataset("P3ps/Cross_ner", split="test")
labels = train_dataset.features["ner_tags"].feature.names
# Initialize a SpanMarker model using a pretrained BERT-style encoder
model_name = "bert-base-cased"
model = SpanMarkerModel.from_pretrained(
model_name,
labels=labels,
# SpanMarker hyperparameters:
model_max_length=256,
marker_max_length=128,
entity_max_length=8,
)
# Prepare the 🤗 transformers training arguments
args = TrainingArguments(
output_dir=f"models/span_marker_bert_base_cross_ner",
run_name=f"bb_cross_ner",
# Training Hyperparameters:
learning_rate=5e-5,
per_device_train_batch_size=32,
per_device_eval_batch_size=32,
num_train_epochs=3,
weight_decay=0.01,
warmup_ratio=0.1,
bf16=True, # Replace `bf16` with `fp16` if your hardware can't use bf16.
# Other Training parameters
logging_first_step=True,
logging_steps=50,
evaluation_strategy="steps",
save_strategy="steps",
eval_steps=200,
save_total_limit=2,
dataloader_num_workers=2,
)
# Initialize the trainer using our model, training args & dataset, and train
trainer = Trainer(
model=model,
args=args,
train_dataset=train_dataset,
eval_dataset=test_dataset,
)
trainer.train()
trainer.save_model(f"models/span_marker_bert_base_cross_ner/checkpoint-final")
# Compute & save the metrics on the test set
metrics = trainer.evaluate(test_dataset, metric_key_prefix="test")
trainer.save_metrics("test", metrics)
trainer.create_model_card()
if __name__ == "__main__":
main()