bert-file-classifier-v2

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

• Loss: 0.5555
• Accuracy: 0.8655

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: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	73	3.0898	0.1448
No log	2.0	146	2.5365	0.4448
No log	3.0	219	2.0822	0.5241
No log	4.0	292	1.7243	0.5793
No log	5.0	365	1.4901	0.6724
No log	6.0	438	1.2608	0.7138
1.948	7.0	511	1.1093	0.7310
1.948	8.0	584	0.9357	0.7862
1.948	9.0	657	0.8845	0.7862
1.948	10.0	730	0.7613	0.8172
1.948	11.0	803	0.7134	0.8103
1.948	12.0	876	0.6805	0.8345
1.948	13.0	949	0.6432	0.8379
0.5068	14.0	1022	0.6068	0.8621
0.5068	15.0	1095	0.6190	0.8448
0.5068	16.0	1168	0.5663	0.8586
0.5068	17.0	1241	0.5458	0.8586
0.5068	18.0	1314	0.6062	0.8379
0.5068	19.0	1387	0.5615	0.8552
0.5068	20.0	1460	0.6120	0.8414
0.1467	21.0	1533	0.5716	0.8655
0.1467	22.0	1606	0.5603	0.8690
0.1467	23.0	1679	0.5601	0.8655
0.1467	24.0	1752	0.5573	0.8586
0.1467	25.0	1825	0.5555	0.8655

Framework versions

• Transformers 4.42.4
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1