bert-lora-for-author-profiling

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: 0.7931
• Age Acc: 0.5827
• Age Precision: 0.5537
• Age Recall: 0.5827
• Age F1: 0.5258
• Age Precision Macro: 0.5152
• Age Recall Macro: 0.2723
• Age F1 Macro: 0.2861
• Gender Acc: 0.6949
• Gender Precision: 0.6949
• Gender Recall: 0.6949
• Gender F1: 0.6949
• Gender Precision Macro: 0.6948
• Gender Recall Macro: 0.6949
• Gender F1 Macro: 0.6948
• Joint Acc: 0.4110
• Avg Acc: 0.6388
• Avg Precision: 0.6243
• Avg Recall: 0.6388
• Avg F1: 0.6104
• Avg Precision Macro: 0.6050
• Avg Recall Macro: 0.4836
• Avg F1 Macro: 0.4905

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: 9.7145e-05
• train_batch_size: 32
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 64
• 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Age Acc	Age Precision	Age Recall	Age F1	Age Precision Macro	Age Recall Macro	Age F1 Macro	Gender Acc	Gender Precision	Gender Recall	Gender F1	Gender Precision Macro	Gender Recall Macro	Gender F1 Macro	Joint Acc	Avg Acc	Avg Precision	Avg Recall	Avg F1	Avg Precision Macro	Avg Recall Macro	Avg F1 Macro
0.8387	0.5155	5000	0.8252	0.5643	0.5195	0.5643	0.5035	0.4352	0.2432	0.2445	0.6811	0.6813	0.6811	0.6811	0.6812	0.6812	0.6811	0.3875	0.6227	0.6004	0.6227	0.5923	0.5582	0.4622	0.4628
0.816	1.0309	10000	0.8115	0.5727	0.5435	0.5727	0.5103	0.4579	0.2526	0.2578	0.6850	0.6857	0.6850	0.6849	0.6855	0.6853	0.6850	0.3961	0.6289	0.6146	0.6289	0.5976	0.5717	0.4690	0.4714
0.8083	1.5464	15000	0.8012	0.5792	0.5481	0.5792	0.5215	0.5131	0.2669	0.2773	0.6901	0.6904	0.6901	0.6901	0.6903	0.6903	0.6901	0.4052	0.6346	0.6193	0.6346	0.6058	0.6017	0.4786	0.4837
0.806	2.0619	20000	0.7962	0.5810	0.5539	0.5810	0.5224	0.5235	0.2680	0.2803	0.6930	0.6932	0.6930	0.6930	0.6930	0.6931	0.6930	0.4083	0.6370	0.6236	0.6370	0.6077	0.6083	0.4805	0.4866
0.7999	2.5773	25000	0.7931	0.5827	0.5537	0.5827	0.5258	0.5152	0.2723	0.2861	0.6949	0.6949	0.6949	0.6949	0.6948	0.6949	0.6948	0.4110	0.6388	0.6243	0.6388	0.6104	0.6050	0.4836	0.4905

Framework versions

• PEFT 0.17.1
• Transformers 4.56.1
• Pytorch 2.6.0+cu124
• Datasets 4.0.0
• Tokenizers 0.22.0