cultural-bert-base-cased-classifier

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

• Loss: 1.1324
• Accuracy: 0.7433
• F1: 0.7385
• Precision: 0.7363
• Recall: 0.7433

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-06
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
0.6272	1.0	196	0.6521	0.7033	0.6949	0.6949	0.7033
0.5687	2.0	392	0.6339	0.7233	0.7150	0.7302	0.7233
0.4809	3.0	588	0.6047	0.7567	0.7527	0.7541	0.7567
0.3967	4.0	784	0.6087	0.7567	0.7514	0.7571	0.7567
0.3732	5.0	980	0.6366	0.7633	0.7589	0.7571	0.7633
0.2892	6.0	1176	0.7116	0.7533	0.7472	0.7481	0.7533
0.3633	7.0	1372	0.7308	0.7533	0.7456	0.7440	0.7533
0.2847	8.0	1568	0.7624	0.75	0.7422	0.7473	0.75
0.2652	9.0	1764	0.8204	0.7433	0.7372	0.7367	0.7433
0.22	10.0	1960	0.8618	0.7333	0.7250	0.7231	0.7333
0.1766	11.0	2156	0.8491	0.7367	0.7314	0.7311	0.7367
0.1237	12.0	2352	0.9106	0.7533	0.7489	0.7480	0.7533
0.1694	13.0	2548	0.9444	0.7467	0.7443	0.7426	0.7467
0.149	14.0	2744	1.0079	0.7367	0.7307	0.7284	0.7367
0.1163	15.0	2940	1.0527	0.7433	0.7382	0.7360	0.7433
0.1153	16.0	3136	1.0669	0.7433	0.7394	0.7374	0.7433
0.0927	17.0	3332	1.1076	0.74	0.7351	0.7341	0.74
0.0548	18.0	3528	1.1197	0.7467	0.7419	0.7398	0.7467
0.1229	19.0	3724	1.1201	0.7433	0.7385	0.7363	0.7433
0.0661	20.0	3920	1.1324	0.7433	0.7385	0.7363	0.7433

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0
• Datasets 3.5.0
• Tokenizers 0.21.1