msi-nat-mini
This model was trained from scratch on the imagefolder dataset.
It achieves the following results on the evaluation set:
- Loss: 0.8600
- Accuracy: 0.6309
- F1: 0.4763
- Precision: 0.6194
- Recall: 0.3870
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-06
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 10
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy | F1 | Precision | Recall |
|---|
| 0.5496 | 1.0 | 2015 | 0.7573 | 0.5955 | 0.4196 | 0.5559 | 0.3369 |
| 0.4807 | 2.0 | 4031 | 0.7416 | 0.6309 | 0.4981 | 0.6074 | 0.4222 |
| 0.4235 | 3.0 | 6047 | 0.7680 | 0.6325 | 0.5047 | 0.6076 | 0.4317 |
| 0.3879 | 4.0 | 8063 | 0.7875 | 0.6339 | 0.4923 | 0.6179 | 0.4092 |
| 0.3702 | 5.0 | 10078 | 0.7923 | 0.6383 | 0.5128 | 0.6168 | 0.4388 |
| 0.3568 | 6.0 | 12094 | 0.8311 | 0.6313 | 0.4969 | 0.6090 | 0.4197 |
| 0.3661 | 7.0 | 14110 | 0.8345 | 0.6316 | 0.4843 | 0.6166 | 0.3987 |
| 0.354 | 8.0 | 16126 | 0.8501 | 0.6305 | 0.4800 | 0.6162 | 0.3931 |
| 0.3569 | 9.0 | 18141 | 0.8552 | 0.6318 | 0.4809 | 0.6193 | 0.3931 |
| 0.3536 | 10.0 | 20150 | 0.8600 | 0.6309 | 0.4763 | 0.6194 | 0.3870 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.0.1+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0
