biobert-v1.1-finetuned-medmcqa-2024-11-25-T17-01-22

This model is a fine-tuned version of dmis-lab/biobert-v1.1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1793
• Accuracy: 0.5374
• F1: 0.5383

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: 0.000159
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 128
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
1.3839	0.9739	14	1.1338	0.5109	0.5122
1.0754	1.9478	28	1.1150	0.5204	0.5224
0.6423	2.9913	43	1.0494	0.5300	0.5305
0.478	3.9652	57	1.1793	0.5374	0.5383
0.245	4.9391	71	1.4143	0.5372	0.5381
0.1731	5.9826	86	1.7302	0.5317	0.5329
0.1229	6.9565	100	1.8354	0.5326	0.5337
0.0833	8.0	115	1.9010	0.5288	0.5299
0.0735	8.9739	129	2.1489	0.5305	0.5314
0.0647	9.7391	140	2.0799	0.5357	0.5368

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3