swinv2-tiny-patch4-window8-256-dmae-humeda-2

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

• Loss: 0.7928
• Accuracy: 0.7115

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: 40

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	2	1.3469	0.5
No log	2.0	4	1.3200	0.4808
No log	3.0	6	1.3124	0.4808
No log	4.0	8	1.2178	0.5
1.1551	5.0	10	1.0957	0.5769
1.1551	6.0	12	1.0359	0.5769
1.1551	7.0	14	1.0103	0.5962
1.1551	8.0	16	0.9382	0.6538
1.1551	9.0	18	0.8748	0.6346
0.9827	10.0	20	0.8836	0.6154
0.9827	11.0	22	0.8574	0.6154
0.9827	12.0	24	0.8494	0.5962
0.9827	13.0	26	0.8226	0.6154
0.9827	14.0	28	0.8242	0.6346
0.8007	15.0	30	0.8304	0.6154
0.8007	16.0	32	0.8447	0.6538
0.8007	17.0	34	0.8228	0.6923
0.8007	18.0	36	0.7928	0.7115
0.8007	19.0	38	0.7822	0.6731
0.6882	20.0	40	0.7750	0.6538
0.6882	21.0	42	0.7726	0.6538
0.6882	22.0	44	0.7898	0.6731
0.6882	23.0	46	0.8021	0.6731
0.6882	24.0	48	0.7834	0.6923
0.6154	25.0	50	0.7634	0.6731
0.6154	26.0	52	0.7584	0.6923
0.6154	27.0	54	0.7773	0.6538
0.6154	28.0	56	0.7830	0.6538
0.6154	29.0	58	0.7719	0.6538
0.541	30.0	60	0.7603	0.6538
0.541	31.0	62	0.7497	0.6731
0.541	32.0	64	0.7381	0.7115
0.541	33.0	66	0.7275	0.6923
0.541	34.0	68	0.7277	0.6923
0.5163	35.0	70	0.7271	0.6923
0.5163	36.0	72	0.7274	0.6923
0.5163	37.0	74	0.7304	0.6923
0.5163	38.0	76	0.7329	0.6923
0.5163	39.0	78	0.7351	0.6923
0.5183	40.0	80	0.7356	0.6923

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1