bert_sm_cv_defined_4

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

• Loss: 1.8439
• Accuracy: 0.802
• Precision: 0.4824
• Recall: 0.2103
• F1: 0.2929
• D-index: 1.5075

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: 5e-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
• lr_scheduler_warmup_steps: 8000
• num_epochs: 20
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	D-index
No log	1.0	250	0.4642	0.813	0.5833	0.1436	0.2305	1.4993
0.5606	2.0	500	0.4524	0.816	0.6	0.1692	0.264	1.5124
0.5606	3.0	750	0.4448	0.82	0.6154	0.2051	0.3077	1.5303
0.4415	4.0	1000	0.4583	0.819	0.6591	0.1487	0.2427	1.5093
0.4415	5.0	1250	0.4659	0.817	0.5652	0.2667	0.3624	1.5473
0.3644	6.0	1500	0.4927	0.805	0.5	0.3077	0.3810	1.5449
0.3644	7.0	1750	0.5493	0.82	0.6230	0.1949	0.2969	1.5267
0.2821	8.0	2000	0.6262	0.815	0.5893	0.1692	0.2629	1.5110
0.2821	9.0	2250	0.6864	0.798	0.4685	0.2667	0.3399	1.5215
0.2084	10.0	2500	0.8556	0.816	0.5696	0.2308	0.3285	1.5337
0.2084	11.0	2750	0.9422	0.814	0.5542	0.2359	0.3309	1.5327
0.114	12.0	3000	1.0528	0.807	0.5109	0.2410	0.3275	1.5249
0.114	13.0	3250	1.2325	0.802	0.4889	0.3385	0.4000	1.5513
0.0612	14.0	3500	1.3032	0.782	0.4148	0.2872	0.3394	1.5066
0.0612	15.0	3750	1.7304	0.806	0.5102	0.1282	0.2049	1.4843
0.0454	16.0	4000	1.6902	0.797	0.4643	0.2667	0.3388	1.5201
0.0454	17.0	4250	1.6552	0.78	0.4172	0.3231	0.3642	1.5161
0.0379	18.0	4500	1.7798	0.806	0.5054	0.2410	0.3264	1.5236
0.0379	19.0	4750	1.8226	0.816	0.5753	0.2154	0.3134	1.5284
0.0487	20.0	5000	1.8439	0.802	0.4824	0.2103	0.2929	1.5075

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3