finetuned-DSCS24-mitre-distilbert-base-uncased-fill-mask

This model is a fine-tuned version of distilbert-base-uncased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.0279

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: 1e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.7512	1.0	65	2.4057
2.3941	2.0	130	2.2818
2.2615	3.0	195	2.2630
2.2113	4.0	260	2.1359
2.1435	5.0	325	2.1022
2.0719	6.0	390	2.0463
2.0483	7.0	455	2.0830
2.0175	8.0	520	1.9946
1.9778	9.0	585	2.0038
1.9831	10.0	650	1.9502
1.8909	11.0	715	1.9914
1.9602	12.0	780	2.0588
1.9309	13.0	845	2.0038
1.9112	14.0	910	1.9957
1.9197	15.0	975	2.0338

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1