swinv2-tiny-patch4-window8-256-ve-UH

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: 1.0154
• 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: 4e-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	1.6092	0.4038
No log	2.0	4	1.6075	0.4231
No log	3.0	6	1.6037	0.4038
No log	4.0	8	1.5960	0.4038
1.6041	5.0	10	1.5820	0.4038
1.6041	6.0	12	1.5578	0.4038
1.6041	7.0	14	1.5218	0.4038
1.6041	8.0	16	1.4849	0.4038
1.6041	9.0	18	1.4459	0.4038
1.4962	10.0	20	1.4109	0.4038
1.4962	11.0	22	1.3941	0.4038
1.4962	12.0	24	1.3865	0.4038
1.4962	13.0	26	1.3754	0.4038
1.4962	14.0	28	1.3655	0.4038
1.3392	15.0	30	1.3794	0.4038
1.3392	16.0	32	1.3800	0.4038
1.3392	17.0	34	1.3404	0.4038
1.3392	18.0	36	1.3337	0.4038
1.3392	19.0	38	1.3602	0.4038
1.2738	20.0	40	1.3384	0.4038
1.2738	21.0	42	1.3248	0.4038
1.2738	22.0	44	1.2693	0.4038
1.2738	23.0	46	1.2395	0.4038
1.2738	24.0	48	1.2427	0.4038
1.2283	25.0	50	1.2885	0.4038
1.2283	26.0	52	1.2916	0.4038
1.2283	27.0	54	1.2353	0.4038
1.2283	28.0	56	1.2032	0.4038
1.2283	29.0	58	1.2100	0.5577
1.1804	30.0	60	1.2110	0.6154
1.1804	31.0	62	1.1710	0.6346
1.1804	32.0	64	1.1323	0.6154
1.1804	33.0	66	1.1083	0.5962
1.1804	34.0	68	1.0935	0.5962
1.0925	35.0	70	1.0853	0.6346
1.0925	36.0	72	1.0622	0.6731
1.0925	37.0	74	1.0154	0.7115
1.0925	38.0	76	0.9901	0.7115
1.0925	39.0	78	0.9925	0.6923
0.9981	40.0	80	0.9865	0.6731
0.9981	41.0	82	0.9540	0.6731
0.9981	42.0	84	0.9316	0.7115
0.9981	43.0	86	0.9304	0.7115
0.9981	44.0	88	0.9246	0.6923
0.9102	45.0	90	0.8785	0.7115
0.9102	46.0	92	0.8422	0.7115
0.9102	47.0	94	0.8381	0.7115
0.9102	48.0	96	0.8359	0.7115
0.9102	49.0	98	0.8444	0.7115
0.8496	50.0	100	0.8287	0.6731
0.8496	51.0	102	0.7973	0.6923
0.8496	52.0	104	0.7799	0.6923
0.8496	53.0	106	0.7780	0.6923
0.8496	54.0	108	0.7820	0.7115
0.7808	55.0	110	0.7896	0.7115
0.7808	56.0	112	0.7737	0.6923
0.7808	57.0	114	0.7631	0.6731
0.7808	58.0	116	0.7635	0.6538
0.7808	59.0	118	0.7779	0.6538
0.757	60.0	120	0.7990	0.6731
0.757	61.0	122	0.8222	0.6538
0.757	62.0	124	0.8204	0.6538
0.757	63.0	126	0.7964	0.6731
0.757	64.0	128	0.7818	0.6538
0.6919	65.0	130	0.7796	0.6346
0.6919	66.0	132	0.7831	0.6346
0.6919	67.0	134	0.7867	0.6346
0.6919	68.0	136	0.7856	0.6346
0.6919	69.0	138	0.7793	0.6538
0.6722	70.0	140	0.7736	0.6538
0.6722	71.0	142	0.7682	0.6538
0.6722	72.0	144	0.7681	0.6538
0.6722	73.0	146	0.7672	0.6538
0.6722	74.0	148	0.7655	0.6538
0.6642	75.0	150	0.7645	0.6538
0.6642	76.0	152	0.7658	0.6538
0.6642	77.0	154	0.7677	0.6538
0.6642	78.0	156	0.7683	0.6538
0.6642	79.0	158	0.7684	0.6538
0.6491	80.0	160	0.7686	0.6538

Framework versions

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