run1

This model is a fine-tuned version of michiyasunaga/BioLinkBERT-base on the sem_eval_2024_task_2 dataset. It achieves the following results on the evaluation set:

• Loss: 1.2153
• Accuracy: 0.64
• Precision: 0.6583
• Recall: 0.64
• F1: 0.6293

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 16
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1
No log	0.99	53	0.6971	0.515	0.5272	0.515	0.4537
0.7029	2.0	107	0.6899	0.535	0.5413	0.535	0.5166
0.7029	2.99	160	0.6855	0.535	0.5399	0.5350	0.5203
0.6955	4.0	214	0.6698	0.565	0.5686	0.5650	0.5592
0.6955	4.99	267	0.6722	0.57	0.5703	0.5700	0.5696
0.6581	6.0	321	0.6367	0.61	0.6104	0.61	0.6096
0.6581	6.99	374	0.6973	0.58	0.5905	0.58	0.5675
0.5796	8.0	428	0.6925	0.625	0.6348	0.625	0.6180
0.5796	8.99	481	0.7539	0.61	0.6364	0.61	0.5902
0.4636	10.0	535	0.9313	0.575	0.6043	0.575	0.5429
0.4636	10.99	588	0.9028	0.615	0.6227	0.615	0.6089
0.3577	12.0	642	0.8694	0.615	0.6227	0.615	0.6089
0.3577	12.99	695	0.9201	0.635	0.6494	0.635	0.6260
0.3041	14.0	749	0.9186	0.645	0.6583	0.645	0.6374
0.3041	14.99	802	1.1683	0.63	0.6578	0.63	0.6129
0.2344	16.0	856	1.1405	0.625	0.6383	0.625	0.6158
0.2344	16.99	909	1.2451	0.625	0.6474	0.625	0.6102
0.208	18.0	963	1.1640	0.65	0.6671	0.65	0.6408
0.208	18.99	1016	1.2081	0.64	0.6583	0.64	0.6293
0.1757	19.81	1060	1.2153	0.64	0.6583	0.64	0.6293

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.15.0