msi-resnet18-pretrain
This model is a fine-tuned version of microsoft/resnet-18 on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.2966
- Accuracy: 0.9088
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
- 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: 4
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 0.2719 | 1.0 | 1562 | 0.3501 | 0.8891 |
| 0.1611 | 2.0 | 3125 | 0.2813 | 0.9118 |
| 0.166 | 3.0 | 4687 | 0.3161 | 0.8930 |
| 0.1071 | 4.0 | 6248 | 0.2966 | 0.9088 |
Framework versions
- Transformers 4.35.2
- Pytorch 2.0.1+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0