swin-tiny-patch4-window7-224-finetuned-bootcamp

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: 0.8963
• Accuracy: 0.7324

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: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.8889	6	4.2849	0.0047
4.3139	1.9259	13	4.1846	0.0329
4.1651	2.9630	20	4.0585	0.0563
4.1651	4.0	27	3.9527	0.0610
3.9272	4.8889	33	3.8813	0.0610
3.7461	5.9259	40	3.7536	0.0845
3.7461	6.9630	47	3.6486	0.1080
3.5254	8.0	54	3.5603	0.1362
3.3478	8.8889	60	3.4566	0.1362
3.3478	9.9259	67	3.2986	0.1502
3.0423	10.9630	74	3.2166	0.1549
2.7931	12.0	81	3.0203	0.2160
2.7931	12.8889	87	2.8991	0.2911
2.541	13.9259	94	2.7941	0.2911
2.3487	14.9630	101	2.7337	0.2911
2.3487	16.0	108	2.5401	0.3662
2.1043	16.8889	114	2.5088	0.3803
1.8892	17.9259	121	2.3596	0.4131
1.8892	18.9630	128	2.3180	0.4178
1.7167	20.0	135	2.1820	0.4272
1.5748	20.8889	141	2.0547	0.4413
1.5748	21.9259	148	1.9472	0.4930
1.4052	22.9630	155	1.9053	0.4883
1.2535	24.0	162	1.8179	0.5117
1.2535	24.8889	168	1.7600	0.5305
1.1687	25.9259	175	1.6922	0.5493
1.0719	26.9630	182	1.6076	0.5587
1.0719	28.0	189	1.5316	0.5587
1.0577	28.8889	195	1.5365	0.5775
0.9558	29.9259	202	1.4488	0.6291
0.9558	30.9630	209	1.4185	0.6150
0.8771	32.0	216	1.3906	0.6056
0.8146	32.8889	222	1.3828	0.6150
0.8146	33.9259	229	1.3927	0.5822
0.8228	34.9630	236	1.3036	0.6385
0.6878	36.0	243	1.2240	0.6808
0.6878	36.8889	249	1.2388	0.6714
0.6471	37.9259	256	1.1345	0.6808
0.6102	38.9630	263	1.1815	0.6573
0.6599	40.0	270	1.1720	0.6526
0.6599	40.8889	276	1.1336	0.6526
0.5742	41.9259	283	1.0863	0.6714
0.5478	42.9630	290	1.0910	0.6714
0.5478	44.0	297	1.0746	0.6620
0.557	44.8889	303	1.0724	0.6808
0.5753	45.9259	310	1.0108	0.7136
0.5753	46.9630	317	1.1296	0.6432
0.5325	48.0	324	1.0361	0.6901
0.4349	48.8889	330	1.0237	0.6995
0.4349	49.9259	337	0.9790	0.7183
0.447	50.9630	344	1.0409	0.6808
0.4502	52.0	351	1.0467	0.6714
0.4502	52.8889	357	0.9773	0.7183
0.4345	53.9259	364	0.9931	0.6808
0.4557	54.9630	371	0.9685	0.7136
0.4557	56.0	378	0.9547	0.7371
0.4109	56.8889	384	1.0015	0.6948
0.4406	57.9259	391	0.9410	0.7230
0.4406	58.9630	398	0.9765	0.6808
0.4039	60.0	405	0.9505	0.7089
0.396	60.8889	411	0.9539	0.7183
0.396	61.9259	418	1.0391	0.6761
0.3958	62.9630	425	0.9576	0.7136
0.3763	64.0	432	0.9380	0.7230
0.3763	64.8889	438	0.9363	0.7277
0.3985	65.9259	445	0.9400	0.7089
0.3701	66.9630	452	0.9769	0.7183
0.3701	68.0	459	0.9604	0.7277
0.3729	68.8889	465	0.9883	0.7089
0.3958	69.9259	472	0.9516	0.7277
0.3958	70.9630	479	0.9252	0.7183
0.359	72.0	486	0.9196	0.7136
0.362	72.8889	492	0.9104	0.7230
0.362	73.9259	499	0.9255	0.7136
0.353	74.9630	506	0.9359	0.7089
0.345	76.0	513	0.9274	0.7230
0.345	76.8889	519	0.9206	0.7371
0.3414	77.9259	526	0.9229	0.7277
0.3298	78.9630	533	0.9102	0.7418
0.3394	80.0	540	0.8955	0.7512
0.3394	80.8889	546	0.8956	0.7371
0.3384	81.9259	553	0.8927	0.7277
0.3164	82.9630	560	0.8885	0.7418
0.3164	84.0	567	0.8941	0.7371
0.3055	84.8889	573	0.8963	0.7418
0.3355	85.9259	580	0.8992	0.7324
0.3355	86.9630	587	0.8988	0.7324
0.3101	88.0	594	0.8969	0.7324
0.3218	88.8889	600	0.8963	0.7324

Framework versions

• Transformers 4.42.3
• Pytorch 2.3.1
• Datasets 2.20.0
• Tokenizers 0.19.1