SpanBERT large fine-tuned on TACRED

SpanBERT created by Facebook Research and fine-tuned on TACRED dataset by them

Details of SpanBERT

SpanBERT: Improving Pre-training by Representing and Predicting Spans

Dataset 📚

TACRED A large-scale relation extraction dataset with 106k+ examples over 42 TAC KBP relation types.

Model fine-tuning 🏋️‍

You can get the fine-tuning script here

python code/run_tacred.py \
  --do_train \
  --do_eval \
  --data_dir <TACRED_DATA_DIR> \
  --model spanbert-large-cased \
  --train_batch_size 32 \
  --eval_batch_size 32 \
  --learning_rate 2e-5 \
  --num_train_epochs 10 \
  --max_seq_length 128 \
  --output_dir tacred_dir \
  --fp16

Results Comparison 📝

	SQuAD 1.1	SQuAD 2.0	Coref	TACRED
	F1	F1	avg. F1	F1
BERT (base)	88.5*	76.5*	73.1	67.7
SpanBERT (base)	92.4*	83.6*	77.4	68.2
BERT (large)	91.3	83.3	77.1	66.4
SpanBERT (large)	94.6	88.7	79.6	70.8 (this one)

Note: The numbers marked as * are evaluated on the development sets because those models were not submitted to the official SQuAD leaderboard. All the other numbers are test numbers.

Created by Manuel Romero/@mrm8488

Made with ♥ in Spain