MsIssuesBERT

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: nan
• Ethnic Boundaries F1: 0.9313
• Ethnic Boundaries Accuracy: 0.9363
• Economic Inequality F1: 0.8031
• Economic Inequality Accuracy: 0.8123
• Economic Policy Benefits F1: 0.8269
• Economic Policy Benefits Accuracy: 0.8485
• Religion Ethnic Identity F1: 0.8491
• Religion Ethnic Identity Accuracy: 0.8588
• Language Policy F1: 0.6336
• Language Policy Accuracy: 0.7059
• Mother Tongue Education F1: 0.8370
• Mother Tongue Education Accuracy: 0.8889
• Overall F1: 0.8135
• Overall Accuracy: 0.8418

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: 4.452845612911518e-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: 964
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Ethnic Boundaries F1	Ethnic Boundaries Accuracy	Economic Inequality F1	Economic Inequality Accuracy	Economic Policy Benefits F1	Economic Policy Benefits Accuracy	Religion Ethnic Identity F1	Religion Ethnic Identity Accuracy	Language Policy F1	Language Policy Accuracy	Mother Tongue Education F1	Mother Tongue Education Accuracy	Overall F1	Overall Accuracy
0.0242	1.0	1000	nan	0.9199	0.9461	0.6796	0.7771	0.7411	0.8215	0.7662	0.8395	0.5459	0.6765	0.6806	0.7778	0.7222	0.8064
0.092	2.0	2000	nan	0.9393	0.9444	0.7938	0.8023	0.7996	0.8316	0.8412	0.8569	0.6336	0.7059	0.8370	0.8889	0.8074	0.8383
0.083	3.0	3000	nan	0.9323	0.9395	0.8053	0.8249	0.8170	0.8519	0.8419	0.8588	0.6071	0.7059	0.8370	0.8889	0.8068	0.8450
1.6647	4.0	4000	nan	0.9298	0.9297	0.8046	0.8098	0.8367	0.8586	0.8604	0.8627	0.6573	0.7353	0.8370	0.8889	0.8210	0.8475
0.0619	5.0	5000	nan	0.9313	0.9363	0.8031	0.8123	0.8269	0.8485	0.8491	0.8588	0.6336	0.7059	0.8370	0.8889	0.8135	0.8418

Framework versions

• Transformers 4.55.4
• Pytorch 2.8.0+cu126
• Datasets 4.0.0
• Tokenizers 0.21.4