README.md

metadata

library_name: transformers
license: apache-2.0
base_model: microsoft/swin-tiny-patch4-window7-224
tags:
  - generated_from_trainer
datasets:
  - imagefolder
metrics:
  - accuracy
model-index:
  - name: swin-tiny-patch4-window7-224-finetuned-eurosat
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: test
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.8071570576540755

swin-tiny-patch4-window7-224-finetuned-eurosat

This model is a fine-tuned version of microsoft/swin-tiny-patch4-window7-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.4697
• Accuracy: 0.8072

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: 5e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 60

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8889	6	0.6376	0.6899
0.6757	1.9259	13	0.6053	0.6938
0.5472	2.9630	20	0.5903	0.7256
0.5472	4.0	27	0.5782	0.7316
0.4628	4.8889	33	0.5979	0.7455
0.4181	5.9259	40	0.5735	0.7614
0.4181	6.9630	47	0.5252	0.7495
0.4079	8.0	54	0.5363	0.7475
0.4102	8.8889	60	0.5289	0.7495
0.4102	9.9259	67	0.5227	0.7535
0.373	10.9630	74	0.4677	0.7773
0.3639	12.0	81	0.4978	0.7813
0.3639	12.8889	87	0.4651	0.7992
0.3779	13.9259	94	0.4738	0.7913
0.3476	14.9630	101	0.4697	0.8072
0.3476	16.0	108	0.4719	0.7952
0.3467	16.8889	114	0.4552	0.7893
0.3496	17.9259	121	0.5186	0.7714
0.3496	18.9630	128	0.4575	0.7952
0.3657	20.0	135	0.4764	0.7793
0.3888	20.8889	141	0.5009	0.7714
0.3888	21.9259	148	0.4673	0.7813
0.3236	22.9630	155	0.4931	0.7753
0.3179	24.0	162	0.4837	0.7654
0.3179	24.8889	168	0.4652	0.7694
0.327	25.9259	175	0.5108	0.7495
0.3253	26.9630	182	0.4424	0.7833
0.3253	28.0	189	0.5622	0.7336
0.3382	28.8889	195	0.5068	0.7694
0.331	29.9259	202	0.4530	0.7694
0.331	30.9630	209	0.5205	0.7316
0.3302	32.0	216	0.4386	0.7853
0.2972	32.8889	222	0.5031	0.7773
0.2972	33.9259	229	0.4909	0.7575
0.3121	34.9630	236	0.4766	0.7793
0.2956	36.0	243	0.5262	0.7416
0.2956	36.8889	249	0.5374	0.7316
0.2947	37.9259	256	0.4888	0.7674
0.2662	38.9630	263	0.4881	0.7694
0.2826	40.0	270	0.4669	0.7893
0.2826	40.8889	276	0.4591	0.7972
0.2768	41.9259	283	0.5090	0.7575
0.2836	42.9630	290	0.5250	0.7495
0.2836	44.0	297	0.4748	0.7654
0.2724	44.8889	303	0.4429	0.7833
0.2498	45.9259	310	0.4460	0.7893
0.2498	46.9630	317	0.4722	0.7793
0.2893	48.0	324	0.4799	0.7714
0.2618	48.8889	330	0.4850	0.7714
0.2618	49.9259	337	0.5152	0.7495
0.2664	50.9630	344	0.5347	0.7396
0.27	52.0	351	0.5343	0.7416
0.27	52.8889	357	0.5330	0.7416
0.2539	53.3333	360	0.5320	0.7396

Framework versions

• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.1.0
• Tokenizers 0.19.1