roberta-large_MBTI_I_MBTI_agg_balanced_25

This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.3974
• F1: 0.6820
• Threshold: 0.49
• F1 At 05: 0.6819

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: 16
• eval_batch_size: 16
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 64
• total_eval_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
• lr_scheduler_warmup_steps: 400
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Threshold	F1 At 05
2.7293	0.2825	200	2.6692	0.5648	0.4400	0.5475
2.5669	0.5650	400	2.4746	0.6586	0.48	0.6494
2.4470	0.8475	600	2.3747	0.6705	0.49	0.6654
2.1590	1.1299	800	2.3351	0.6807	0.3400	0.6584
2.0981	1.4124	1000	2.3197	0.6860	0.31	0.6653
1.9872	1.6949	1200	2.4732	0.6810	0.48	0.6797
1.9866	1.9774	1400	2.4751	0.6869	0.47	0.6831
1.5646	2.2599	1600	2.8745	0.6816	0.51	0.6800
1.4286	2.5424	1800	2.9926	0.6868	0.48	0.6842
1.5118	2.8249	2000	2.8057	0.6888	0.5	0.6888
1.0639	3.1073	2200	3.5753	0.6860	0.39	0.6839
1.0155	3.3898	2400	3.6367	0.6826	0.4200	0.6786
1.0504	3.6723	2600	3.3507	0.6872	0.49	0.6868
1.0401	3.9548	2800	3.6559	0.6798	0.51	0.6790
0.8185	4.2373	3000	4.1669	0.6782	0.35	0.6751
0.7157	4.5198	3200	4.5369	0.6810	0.4500	0.6791
0.8082	4.8023	3400	4.3974	0.6820	0.49	0.6819

Framework versions

• Transformers 5.3.0
• Pytorch 2.7.1
• Datasets 4.8.5
• Tokenizers 0.22.2