README.md

metadata

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-PE
    results:
      - task:
          name: Image Classification
          type: image-classification
        dataset:
          name: imagefolder
          type: imagefolder
          config: default
          split: train
          args: default
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5833333333333334

swin-tiny-patch4-window7-224-PE

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: 0.6756
• Accuracy: 0.5833

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: 0.0025
• train_batch_size: 128
• eval_batch_size: 128
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 512
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5675	0.99	20	0.5504	0.7463
0.7158	1.98	40	0.9070	0.5944
0.6498	2.96	60	0.6501	0.6037
0.6405	4.0	81	0.5655	0.7389
0.7003	4.99	101	0.6786	0.5907
0.6857	5.98	121	0.6820	0.5370
0.6933	6.96	141	0.6819	0.5926
0.6795	8.0	162	0.6783	0.5481
0.6872	8.99	182	0.6907	0.5370
0.6942	9.98	202	0.6922	0.5407
0.6945	10.96	222	0.6935	0.4630
0.6936	12.0	243	0.6974	0.4630
0.6935	12.99	263	0.6907	0.5407
0.6925	13.98	283	0.6945	0.4241
0.6927	14.96	303	0.6952	0.4630
0.6921	16.0	324	0.6901	0.5463
0.6937	16.99	344	0.6935	0.4407
0.6933	17.98	364	0.6922	0.5537
0.6929	18.96	384	0.6971	0.4630
0.6919	20.0	405	0.6901	0.5630
0.6903	20.99	425	0.6850	0.5722
0.6892	21.98	445	0.6876	0.5611
0.6846	22.96	465	0.6871	0.5463
0.6841	24.0	486	0.6742	0.5685
0.682	24.99	506	0.6776	0.5741
0.6796	25.98	526	0.6850	0.5407
0.6849	26.96	546	0.6722	0.5907
0.6855	28.0	567	0.6818	0.5648
0.6903	28.99	587	0.7024	0.4685
0.6845	29.98	607	0.6781	0.5630
0.6806	30.96	627	0.6771	0.5778
0.6808	32.0	648	0.6718	0.5833
0.6811	32.99	668	0.6715	0.5833
0.6814	33.98	688	0.6641	0.6370
0.6848	34.96	708	0.6736	0.6111
0.6848	36.0	729	0.6694	0.6259
0.6848	36.99	749	0.6757	0.5907
0.6865	37.98	769	0.6763	0.5667
0.6876	38.96	789	0.6812	0.5889
0.6858	40.0	810	0.6763	0.5926
0.6863	40.99	830	0.6743	0.5981
0.6838	41.98	850	0.6740	0.5796
0.6833	42.96	870	0.6770	0.5611
0.6883	44.0	891	0.6733	0.6037
0.684	44.99	911	0.6730	0.6019
0.6869	45.98	931	0.6731	0.6130
0.6861	46.96	951	0.6752	0.5704
0.686	48.0	972	0.6761	0.5704
0.683	48.99	992	0.6759	0.5722
0.6847	49.38	1000	0.6756	0.5833

Framework versions

• Transformers 4.35.2
• Pytorch 2.0.1+cu117
• Datasets 2.15.0
• Tokenizers 0.15.0