roberta-large_MBTI_F_MBTI_agg_balanced_50_1e-05_0.05
This model was trained from scratch on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 3.2626
- F1: 0.7462
- Threshold: 0.76
- F1 At 05: 0.7450
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 |
|---|---|---|---|---|---|---|
| 2.3707 | 0.3175 | 200 | 2.3279 | 0.6838 | 0.48 | 0.6813 |
| 2.2372 | 0.6349 | 400 | 2.2484 | 0.7109 | 0.4100 | 0.6931 |
| 2.0317 | 0.9524 | 600 | 2.0265 | 0.7463 | 0.47 | 0.7433 |
| 1.7655 | 1.2698 | 800 | 2.0405 | 0.7529 | 0.38 | 0.7457 |
| 1.6984 | 1.5873 | 1000 | 2.1494 | 0.7462 | 0.55 | 0.7370 |
| 1.6636 | 1.9048 | 1200 | 2.1611 | 0.7507 | 0.63 | 0.7423 |
| 1.2216 | 2.2222 | 1400 | 2.3145 | 0.7531 | 0.64 | 0.7491 |
| 1.1912 | 2.5397 | 1600 | 2.5443 | 0.7541 | 0.6 | 0.7504 |
| 1.1460 | 2.8571 | 1800 | 2.4287 | 0.7537 | 0.33 | 0.7453 |
| 0.8011 | 3.1746 | 2000 | 2.7473 | 0.7475 | 0.65 | 0.7419 |
| 0.8214 | 3.4921 | 2200 | 2.7369 | 0.7487 | 0.51 | 0.7484 |
| 0.7635 | 3.8095 | 2400 | 2.8950 | 0.7505 | 0.66 | 0.7449 |
| 0.5983 | 4.1270 | 2600 | 3.2703 | 0.7509 | 0.8 | 0.7395 |
| 0.5477 | 4.4444 | 2800 | 3.2470 | 0.7468 | 0.67 | 0.7447 |
| 0.5415 | 4.7619 | 3000 | 3.2626 | 0.7462 | 0.76 | 0.7450 |
Framework versions
- Transformers 5.3.0
- Pytorch 2.7.1
- Datasets 4.8.5
- Tokenizers 0.22.2