muhtasham/bert-small-finetuned-eoir_privacy-longer10

bert_uncased_L-4_H-512_A-8-finetuned-eoir_privacy-longer

This model is a fine-tuned version of google/bert_uncased_L-4_H-512_A-8 on the eoir_privacy dataset. It achieves the following results on the evaluation set:

• Loss: 0.1738
• Accuracy: 0.9476
• F1: 0.8805

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
No log	1.0	63	0.2004	0.9261	0.8298
No log	2.0	126	0.2042	0.9297	0.8435
No log	3.0	189	0.1888	0.9362	0.8558
No log	4.0	252	0.1879	0.9383	0.8532
No log	5.0	315	0.1715	0.9462	0.8788
No log	6.0	378	0.1761	0.9448	0.8706
No log	7.0	441	0.1730	0.9455	0.8774
0.1078	8.0	504	0.1753	0.9448	0.8772
0.1078	9.0	567	0.1721	0.9469	0.8791
0.1078	10.0	630	0.1738	0.9476	0.8805

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.1+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1