swinv2-tiny-patch4-window8-256-ve-UH2

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: 0.7448
• Accuracy: 0.7308

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: 80

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	2	1.6091	0.4038
No log	2.0	4	1.6070	0.4231
No log	3.0	6	1.6017	0.4038
No log	4.0	8	1.5911	0.4038
1.6022	5.0	10	1.5707	0.4038
1.6022	6.0	12	1.5355	0.4038
1.6022	7.0	14	1.4951	0.4038
1.6022	8.0	16	1.4528	0.4038
1.6022	9.0	18	1.4096	0.4038
1.4645	10.0	20	1.3820	0.4038
1.4645	11.0	22	1.4055	0.4038
1.4645	12.0	24	1.3765	0.4038
1.4645	13.0	26	1.3820	0.4038
1.4645	14.0	28	1.3712	0.4038
1.3172	15.0	30	1.3546	0.4038
1.3172	16.0	32	1.3637	0.4038
1.3172	17.0	34	1.3646	0.4038
1.3172	18.0	36	1.3271	0.4038
1.3172	19.0	38	1.3084	0.4038
1.2549	20.0	40	1.3402	0.4038
1.2549	21.0	42	1.3550	0.4038
1.2549	22.0	44	1.2677	0.4038
1.2549	23.0	46	1.2093	0.4038
1.2549	24.0	48	1.2040	0.4231
1.2092	25.0	50	1.2963	0.4231
1.2092	26.0	52	1.2917	0.4808
1.2092	27.0	54	1.1798	0.5769
1.2092	28.0	56	1.1047	0.6346
1.2092	29.0	58	1.0923	0.6731
1.1321	30.0	60	1.1066	0.6538
1.1321	31.0	62	1.0874	0.6538
1.1321	32.0	64	1.0548	0.6731
1.1321	33.0	66	1.0012	0.6538
1.1321	34.0	68	0.9641	0.6923
1.0022	35.0	70	0.9796	0.6538
1.0022	36.0	72	0.9631	0.6538
1.0022	37.0	74	0.9040	0.6731
1.0022	38.0	76	0.8731	0.6923
1.0022	39.0	78	0.8960	0.6731
0.8941	40.0	80	0.9133	0.6538
0.8941	41.0	82	0.8507	0.6923
0.8941	42.0	84	0.8064	0.7115
0.8941	43.0	86	0.8075	0.7115
0.8941	44.0	88	0.8486	0.6923
0.7866	45.0	90	0.8075	0.6923
0.7866	46.0	92	0.7496	0.6731
0.7866	47.0	94	0.7431	0.6731
0.7866	48.0	96	0.7442	0.6731
0.7866	49.0	98	0.7735	0.6923
0.7281	50.0	100	0.7751	0.6923
0.7281	51.0	102	0.7370	0.6923
0.7281	52.0	104	0.7230	0.6923
0.7281	53.0	106	0.7314	0.6923
0.7281	54.0	108	0.7498	0.6731
0.6725	55.0	110	0.7557	0.6731
0.6725	56.0	112	0.7314	0.7115
0.6725	57.0	114	0.7334	0.7115
0.6725	58.0	116	0.7375	0.7115
0.6725	59.0	118	0.7434	0.6923
0.6526	60.0	120	0.7548	0.6731
0.6526	61.0	122	0.7813	0.7115
0.6526	62.0	124	0.7722	0.6923
0.6526	63.0	126	0.7469	0.6923
0.6526	64.0	128	0.7402	0.6731
0.5915	65.0	130	0.7448	0.7308
0.5915	66.0	132	0.7467	0.6923
0.5915	67.0	134	0.7496	0.6731
0.5915	68.0	136	0.7518	0.7308
0.5915	69.0	138	0.7453	0.7115
0.578	70.0	140	0.7385	0.6923
0.578	71.0	142	0.7411	0.6731
0.578	72.0	144	0.7442	0.6731
0.578	73.0	146	0.7440	0.6731
0.578	74.0	148	0.7428	0.6923
0.5826	75.0	150	0.7414	0.6923
0.5826	76.0	152	0.7416	0.6923
0.5826	77.0	154	0.7414	0.6923
0.5826	78.0	156	0.7413	0.6731
0.5826	79.0	158	0.7413	0.6731
0.5586	80.0	160	0.7415	0.6923

Framework versions

• Transformers 4.36.2
• Pytorch 2.1.2+cu118
• Datasets 2.16.1
• Tokenizers 0.15.0