distilbert-no-aug-pubmed

This model is a fine-tuned version of distilbert/distilbert-base-multilingual-cased on the pietrolesci/pubmed-20k-rct dataset. It achieves the following results on the evaluation set:

• Loss: 0.7624
• F1: 0.6499

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: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	32	1.3359	0.4524
No log	2.0	64	0.9877	0.5552
No log	3.0	96	0.7930	0.6458
No log	4.0	128	0.7555	0.6404
No log	5.0	160	0.7209	0.6516
No log	6.0	192	0.7981	0.6238
No log	7.0	224	0.7684	0.6489
No log	8.0	256	0.7444	0.6535
No log	9.0	288	0.7954	0.6307
No log	10.0	320	0.7624	0.6499

Framework versions

• Transformers 4.57.1
• Pytorch 2.9.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1