vit-gpt-NWPU-captioning
This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.6254
- Accuracy: 85.1
- Bleu-1: 0.8526
- Bleu-2: 0.7615
- Bleu-3: 0.6873
- Bleu-4: 0.6274
- Meteor: 0.7682
- Rouge-l: 0.7527
- Cider: 1.7556
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: 0.0001
- train_batch_size: 64
- eval_batch_size: 64
- seed: 50
- 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: 1024
- num_epochs: 128
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | Bleu-1 | Bleu-2 | Bleu-3 | Bleu-4 | Meteor | Rouge-l | Cider |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.8241 | 1.0 | 1969 | 0.5023 | 85.06 | 0.8202 | 0.7130 | 0.6327 | 0.5699 | 0.7179 | 0.7020 | 1.6309 |
| 0.4855 | 2.0 | 3938 | 0.4788 | 84.88 | 0.8371 | 0.7374 | 0.6566 | 0.5926 | 0.7398 | 0.7235 | 1.6648 |
| 0.4499 | 3.0 | 5907 | 0.4754 | 85.1 | 0.8476 | 0.7561 | 0.6821 | 0.6236 | 0.7543 | 0.7413 | 1.7213 |
| 0.4260 | 4.0 | 7876 | 0.4710 | 84.49 | 0.8477 | 0.7575 | 0.6852 | 0.6268 | 0.7705 | 0.7533 | 1.7417 |
| 0.4066 | 5.0 | 9845 | 0.4763 | 85.14 | 0.8618 | 0.7728 | 0.7017 | 0.6441 | 0.7703 | 0.7583 | 1.8203 |
| 0.3883 | 6.0 | 11814 | 0.4791 | 85.28 | 0.8654 | 0.7775 | 0.7076 | 0.6513 | 0.7669 | 0.7625 | 1.8222 |
| 0.3686 | 7.0 | 13783 | 0.4883 | 84.91 | 0.8567 | 0.7669 | 0.6941 | 0.6351 | 0.7770 | 0.7606 | 1.7645 |
| 0.3499 | 8.0 | 15752 | 0.4983 | 85.09 | 0.8609 | 0.7744 | 0.7048 | 0.6480 | 0.7752 | 0.7632 | 1.8206 |
| 0.3313 | 9.0 | 17721 | 0.5119 | 84.99 | 0.8604 | 0.7729 | 0.7027 | 0.6456 | 0.7758 | 0.7621 | 1.8067 |
| 0.3115 | 10.0 | 19690 | 0.5315 | 85.26 | 0.8652 | 0.7799 | 0.7112 | 0.6549 | 0.7835 | 0.7700 | 1.8623 |
| 0.2940 | 11.0 | 21659 | 0.5515 | 85.11 | 0.8549 | 0.7636 | 0.6921 | 0.6345 | 0.7664 | 0.7537 | 1.7845 |
| 0.2775 | 12.0 | 23628 | 0.5757 | 85.1 | 0.8508 | 0.7575 | 0.6832 | 0.6235 | 0.7588 | 0.7468 | 1.7401 |
| 0.2458 | 13.0 | 25597 | 0.5890 | 85.13 | 0.8510 | 0.7551 | 0.6800 | 0.6201 | 0.7577 | 0.7445 | 1.7500 |
| 0.2295 | 14.0 | 27566 | 0.6115 | 85.25 | 0.8500 | 0.7584 | 0.6859 | 0.6272 | 0.7593 | 0.7503 | 1.7433 |
| 0.2204 | 15.0 | 29535 | 0.6254 | 85.1 | 0.8526 | 0.7615 | 0.6873 | 0.6274 | 0.7682 | 0.7527 | 1.7556 |
Framework versions
- Transformers 5.12.1
- Pytorch 2.12.1+cu130
- Datasets 5.0.0
- Tokenizers 0.22.2