roberta-large_MBTI_I_MBTI_agg_balanced_75

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

• Loss: 3.9016
• F1: 0.6801
• Threshold: 0.49
• F1 At 05: 0.6791

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
2.7287	0.3509	200	2.7492	0.5840	0.51	0.5812
2.5910	0.7018	400	2.5378	0.6425	0.4400	0.5598
2.1510	1.0526	600	2.4344	0.6753	0.4400	0.6593
1.9984	1.4035	800	2.4543	0.6859	0.47	0.6792
1.9386	1.7544	1000	2.3035	0.6940	0.4600	0.6847
1.4981	2.1053	1200	2.5572	0.6912	0.4400	0.6846
1.4341	2.4561	1400	2.5054	0.6929	0.47	0.6866
1.4262	2.8070	1600	2.7249	0.6915	0.4200	0.6888
1.0158	3.1579	1800	3.1633	0.6895	0.4400	0.6845
0.9583	3.5088	2000	3.4025	0.6751	0.36	0.6680
0.9158	3.8596	2200	3.1368	0.6871	0.36	0.6797
0.6699	4.2105	2400	4.0544	0.6762	0.53	0.6762
0.6745	4.5614	2600	3.7200	0.6816	0.38	0.6793
0.7178	4.9123	2800	3.9016	0.6801	0.49	0.6791

Framework versions

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