favs_filter_classification_v2

This model is a fine-tuned version of bert-base-cased on the filter_v2 dataset. It achieves the following results on the evaluation set:

• Loss: 0.2016
• F1: 0.9762
• Roc Auc: 0.9844
• Accuracy: 0.9545

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: 1.5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.6596	1.0	16	0.6086	0.2687	0.5474	0.0
0.5448	2.0	32	0.5354	0.3824	0.6063	0.0
0.5106	3.0	48	0.4874	0.4444	0.6382	0.0455
0.4353	4.0	64	0.4301	0.5352	0.6889	0.1818
0.3699	5.0	80	0.3890	0.6579	0.7640	0.3636
0.349	6.0	96	0.3663	0.6667	0.7633	0.3182
0.3104	7.0	112	0.3327	0.7105	0.7953	0.4545
0.3023	8.0	128	0.2971	0.7733	0.8303	0.5455
0.2676	9.0	144	0.2766	0.8395	0.8861	0.7727
0.2374	10.0	160	0.2541	0.8537	0.8980	0.7727
0.2238	11.0	176	0.2399	0.9024	0.9293	0.8182
0.2084	12.0	192	0.2221	0.9286	0.9531	0.8636
0.2143	13.0	208	0.2138	0.9286	0.9531	0.8636
0.1846	14.0	224	0.2016	0.9762	0.9844	0.9545
0.1812	15.0	240	0.1957	0.9762	0.9844	0.9545
0.1756	16.0	256	0.1881	0.9647	0.9806	0.9091
0.1662	17.0	272	0.1845	0.9762	0.9844	0.9545
0.1715	18.0	288	0.1802	0.9762	0.9844	0.9545
0.1585	19.0	304	0.1782	0.9762	0.9844	0.9545
0.1595	20.0	320	0.1775	0.9762	0.9844	0.9545

Framework versions

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