swiftformer-xs-DMAE
This model is a fine-tuned version of MBZUAI/swiftformer-xs on the imagefolder dataset. It achieves the following results on the evaluation set:
- Loss: 0.7654
- Accuracy: 0.7609
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: 0.005
- 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: 40
Training results
Training Loss | Epoch | Step | Validation Loss | Accuracy |
---|---|---|---|---|
No log | 0.86 | 3 | 1.3706 | 0.5435 |
No log | 2.0 | 7 | 1.1672 | 0.6304 |
1.2937 | 2.86 | 10 | 1.2248 | 0.6087 |
1.2937 | 4.0 | 14 | 2.6986 | 0.3478 |
1.2937 | 4.86 | 17 | 2.6008 | 0.4565 |
1.2355 | 6.0 | 21 | 1.8900 | 0.4130 |
1.2355 | 6.86 | 24 | 1.5958 | 0.5 |
1.2355 | 8.0 | 28 | 1.3847 | 0.5652 |
1.1548 | 8.86 | 31 | 0.8762 | 0.7174 |
1.1548 | 10.0 | 35 | 1.0871 | 0.5435 |
1.1548 | 10.86 | 38 | 0.9779 | 0.5435 |
1.0561 | 12.0 | 42 | 1.2454 | 0.5435 |
1.0561 | 12.86 | 45 | 1.2277 | 0.5652 |
1.0561 | 14.0 | 49 | 0.9527 | 0.6087 |
1.0546 | 14.86 | 52 | 0.8797 | 0.6304 |
1.0546 | 16.0 | 56 | 0.9479 | 0.6304 |
1.0546 | 16.86 | 59 | 0.8696 | 0.6739 |
0.9493 | 18.0 | 63 | 0.9348 | 0.6522 |
0.9493 | 18.86 | 66 | 0.9890 | 0.5435 |
0.9354 | 20.0 | 70 | 0.9073 | 0.5870 |
0.9354 | 20.86 | 73 | 0.8763 | 0.6304 |
0.9354 | 22.0 | 77 | 0.9592 | 0.6522 |
0.8791 | 22.86 | 80 | 0.8940 | 0.6522 |
0.8791 | 24.0 | 84 | 0.8165 | 0.6957 |
0.8791 | 24.86 | 87 | 0.8249 | 0.6957 |
0.8017 | 26.0 | 91 | 0.8946 | 0.6739 |
0.8017 | 26.86 | 94 | 0.8210 | 0.6957 |
0.8017 | 28.0 | 98 | 0.7654 | 0.7609 |
0.8532 | 28.86 | 101 | 0.7492 | 0.7391 |
0.8532 | 30.0 | 105 | 0.7725 | 0.6957 |
0.8532 | 30.86 | 108 | 0.7932 | 0.7174 |
0.8205 | 32.0 | 112 | 0.8129 | 0.7391 |
0.8205 | 32.86 | 115 | 0.8089 | 0.6957 |
0.8205 | 34.0 | 119 | 0.8151 | 0.6957 |
0.8112 | 34.29 | 120 | 0.8074 | 0.6957 |
Framework versions
- Transformers 4.36.2
- Pytorch 2.1.2+cu118
- Datasets 2.16.1
- Tokenizers 0.15.0
- Downloads last month
- 23
Model tree for Augusto777/swiftformer-xs-DMAE
Base model
MBZUAI/swiftformer-xsEvaluation results
- Accuracy on imagefoldervalidation set self-reported0.761