bert-base-uncased_12112024T130600

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

• Loss: 1.2744
• F1: 0.4792
• Learning Rate: 0.0

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 600
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Rate
No log	0.9942	86	1.8801	0.0859	0.0000
No log	2.0	173	1.8639	0.1526	0.0000
No log	2.9942	259	1.7658	0.3384	0.0000
No log	4.0	346	1.5763	0.4250	0.0000
No log	4.9942	432	1.3998	0.4554	0.0000
1.5137	6.0	519	1.3179	0.4529	0.0000
1.5137	6.9942	605	1.3344	0.4610	1e-05
1.5137	8.0	692	1.2744	0.4792	0.0000
1.5137	8.9942	778	1.3345	0.4778	0.0000
1.5137	10.0	865	1.3757	0.4794	0.0000
1.5137	10.9942	951	1.4471	0.4857	0.0000
0.6472	12.0	1038	1.5127	0.4885	0.0000
0.6472	12.9942	1124	1.6826	0.4786	0.0000
0.6472	14.0	1211	1.6606	0.4938	0.0000
0.6472	14.9942	1297	1.8626	0.4945	0.0000
0.6472	16.0	1384	1.9157	0.4965	0.0000
0.6472	16.9942	1470	1.9586	0.5034	6e-06
0.179	18.0	1557	1.9809	0.5039	0.0000
0.179	18.9942	1643	2.1634	0.5057	0.0000
0.179	20.0	1730	2.3241	0.4995	0.0000
0.179	20.9942	1816	2.3038	0.5037	0.0000
0.179	22.0	1903	2.3403	0.5091	0.0000
0.179	22.9942	1989	2.3770	0.5044	0.0000
0.0501	24.0	2076	2.5511	0.4936	0.0000
0.0501	24.9942	2162	2.5529	0.4976	0.0000
0.0501	26.0	2249	2.5132	0.5041	7e-07
0.0501	26.9942	2335	2.5330	0.5049	4e-07
0.0501	28.0	2422	2.5368	0.4989	2e-07
0.0248	28.9942	2508	2.5301	0.4973	0.0
0.0248	29.8266	2580	2.5316	0.4973	0.0

Framework versions

• Transformers 4.44.2
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.19.1