swinv2-tiny-patch4-window8-256-DMAE-4e-3

This model is a fine-tuned version of microsoft/swinv2-tiny-patch4-window8-256 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.7507
• Accuracy: 0.7391

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8571	3	1.3959	0.3043
No log	1.7857	6	1.2662	0.3913
No log	2.7143	9	1.1960	0.4783
1.3226	3.9286	13	1.1950	0.4565
1.3226	4.8571	16	1.1891	0.4783
1.3226	5.7857	19	1.1898	0.4783
1.1833	6.7143	22	1.1824	0.5435
1.1833	7.9286	26	1.1618	0.5217
1.1833	8.8571	29	1.1359	0.5652
1.1384	9.7857	32	1.0974	0.5870
1.1384	10.7143	35	1.0524	0.5870
1.1384	11.9286	39	1.0083	0.6957
1.0628	12.8571	42	0.9696	0.6739
1.0628	13.7857	45	0.9369	0.6739
1.0628	14.7143	48	0.8825	0.7174
1.0069	15.9286	52	0.8396	0.6957
1.0069	16.8571	55	0.8267	0.7174
1.0069	17.7857	58	0.8275	0.7174
0.9339	18.7143	61	0.8255	0.7174
0.9339	19.9286	65	0.7899	0.7174
0.9339	20.8571	68	0.7604	0.7174
0.905	21.7857	71	0.7442	0.6957
0.905	22.7143	74	0.7361	0.7391
0.905	23.9286	78	0.7598	0.6957
0.8465	24.8571	81	0.7650	0.7174
0.8465	25.7857	84	0.7631	0.7391
0.8465	26.7143	87	0.7561	0.7174
0.8363	27.9286	91	0.7494	0.6957
0.8363	28.8571	94	0.7539	0.7174
0.8363	29.7857	97	0.7497	0.7174
0.7751	30.7143	100	0.7477	0.7174
0.7751	31.9286	104	0.7463	0.7609
0.7751	32.8571	107	0.7507	0.7609
0.7843	33.7857	110	0.7534	0.7391
0.7843	34.7143	113	0.7542	0.7391
0.7843	35.9286	117	0.7519	0.7391
0.7435	36.8571	120	0.7507	0.7391

Framework versions

• Transformers 4.46.2
• Pytorch 2.5.1+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3