bert_base_for_whole_train_result_Spam-Ham2_4

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

• Loss: 0.0396
• Accuracy: 0.994
• F1: 0.9944

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: 0.0001
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 4096
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
0.71	6.8817	50	0.5082	0.865	0.8667
0.2172	13.7634	100	0.0715	0.976	0.9771
0.0161	20.6452	150	0.0341	0.9915	0.9920
0.0045	27.5269	200	0.0270	0.994	0.9944
0.0038	34.4086	250	0.0433	0.99	0.9905
0.0013	41.2903	300	0.0333	0.9945	0.9948
0.0006	48.1720	350	0.0331	0.9955	0.9958
0.0008	55.0538	400	0.0435	0.99	0.9906
0.0002	61.9355	450	0.0375	0.993	0.9934
0.0002	68.8172	500	0.0434	0.993	0.9934
0.0008	75.6989	550	0.0315	0.9945	0.9948
0.0002	82.5806	600	0.0336	0.995	0.9953
0.0005	89.4624	650	0.0525	0.992	0.9924
0.0006	96.3441	700	0.0396	0.994	0.9944

Framework versions

• Transformers 4.40.0
• Pytorch 2.4.1+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1