MyPoliBERT-HITL

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: 0.2658
• Democracy F1: 0.8047
• Democracy Accuracy: 0.8399
• Economy F1: 0.8967
• Economy Accuracy: 0.9108
• Race F1: 0.9393
• Race Accuracy: 0.9475
• Leadership F1: 0.7616
• Leadership Accuracy: 0.8084
• Development F1: 0.9365
• Development Accuracy: 0.9475
• Corruption F1: 0.9298
• Corruption Accuracy: 0.9449
• Stability F1: 0.8689
• Stability Accuracy: 0.8924
• Safety F1: 0.9170
• Safety Accuracy: 0.9265
• Administration F1: 0.8177
• Administration Accuracy: 0.8635
• Education F1: 0.9820
• Education Accuracy: 0.9843
• Religion F1: 0.9323
• Religion Accuracy: 0.9423
• Environment F1: 0.9823
• Environment Accuracy: 0.9843
• Overall F1: 0.8974
• Overall Accuracy: 0.9160

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: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Democracy F1	Democracy Accuracy	Economy F1	Economy Accuracy	Race F1	Race Accuracy	Leadership F1	Leadership Accuracy	Development F1	Development Accuracy	Corruption F1	Corruption Accuracy	Stability F1	Stability Accuracy	Safety F1	Safety Accuracy	Administration F1	Administration Accuracy	Education F1	Education Accuracy	Religion F1	Religion Accuracy	Environment F1	Environment Accuracy	Overall F1	Overall Accuracy
No log	1.0	96	0.7728	0.7594	0.8346	0.8568	0.9029	0.8912	0.9265	0.6722	0.7717	0.9065	0.9370	0.8759	0.9160	0.8341	0.8871	0.8078	0.8688	0.7891	0.8556	0.9608	0.9738	0.8720	0.9134	0.9297	0.9528	0.8463	0.8950
1.3991	2.0	192	0.4467	0.7594	0.8346	0.8568	0.9029	0.8912	0.9265	0.6722	0.7717	0.9065	0.9370	0.8759	0.9160	0.8341	0.8871	0.8078	0.8688	0.7891	0.8556	0.9608	0.9738	0.8720	0.9134	0.9297	0.9528	0.8463	0.8950
0.5431	3.0	288	0.3804	0.7594	0.8346	0.8568	0.9029	0.8912	0.9265	0.6851	0.7769	0.9065	0.9370	0.8759	0.9160	0.8341	0.8871	0.8078	0.8688	0.7891	0.8556	0.9785	0.9843	0.8720	0.9134	0.9297	0.9528	0.8488	0.8963
0.4275	4.0	384	0.3278	0.7581	0.8320	0.8715	0.9055	0.8912	0.9265	0.7398	0.7874	0.9065	0.9370	0.8988	0.9239	0.8453	0.8740	0.8794	0.9029	0.7891	0.8556	0.9790	0.9843	0.8893	0.9213	0.9773	0.9816	0.8688	0.9027
0.35	5.0	480	0.3028	0.7711	0.8136	0.8820	0.9081	0.9250	0.9396	0.7487	0.7979	0.9117	0.9396	0.9167	0.9318	0.8570	0.8898	0.9019	0.9134	0.7891	0.8556	0.9693	0.9711	0.9093	0.9291	0.9709	0.9764	0.8794	0.9055
0.2846	6.0	576	0.2856	0.7767	0.8215	0.9040	0.9186	0.9183	0.9344	0.7716	0.8189	0.9240	0.9449	0.9140	0.9344	0.8518	0.8688	0.9053	0.9186	0.7891	0.8556	0.9771	0.9816	0.9317	0.9423	0.9758	0.9790	0.8866	0.9099
0.23	7.0	672	0.2759	0.7920	0.8373	0.8959	0.9134	0.9312	0.9423	0.7453	0.8058	0.9299	0.9475	0.9258	0.9423	0.8516	0.8871	0.9061	0.9213	0.7962	0.8583	0.9771	0.9816	0.9391	0.9475	0.9818	0.9843	0.8893	0.9140
0.1845	8.0	768	0.2677	0.7986	0.8425	0.8956	0.9081	0.9380	0.9449	0.7611	0.8110	0.9325	0.9475	0.9298	0.9449	0.8590	0.8898	0.9148	0.9239	0.8095	0.8556	0.9771	0.9816	0.9305	0.9423	0.9830	0.9843	0.8941	0.9147
0.1643	9.0	864	0.2664	0.8099	0.8451	0.8909	0.9081	0.9367	0.9449	0.7676	0.8163	0.9370	0.9475	0.9298	0.9449	0.8643	0.8898	0.9125	0.9213	0.8050	0.8530	0.9820	0.9843	0.9344	0.9449	0.9793	0.9816	0.8958	0.9151
0.1353	10.0	960	0.2658	0.8047	0.8399	0.8967	0.9108	0.9393	0.9475	0.7616	0.8084	0.9365	0.9475	0.9298	0.9449	0.8689	0.8924	0.9170	0.9265	0.8177	0.8635	0.9820	0.9843	0.9323	0.9423	0.9823	0.9843	0.8974	0.9160

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.0