sashes_model

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.3784
• Accuracy: 0.8760

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: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 112

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.9697	8	2.2973	0.1434
2.2994	1.9394	16	2.2717	0.1957
2.2791	2.9091	24	2.2377	0.2287
2.2378	4.0	33	2.1866	0.3178
2.1604	4.9697	41	2.1096	0.3934
2.1604	5.9394	49	2.0257	0.4322
2.0801	6.9091	57	1.9312	0.4264
1.9587	8.0	66	1.7939	0.4942
1.821	8.9697	74	1.6869	0.5465
1.6903	9.9394	82	1.6025	0.5736
1.5687	10.9091	90	1.4849	0.6202
1.5687	12.0	99	1.4674	0.5407
1.4183	12.9697	107	1.3539	0.6163
1.3907	13.9394	115	1.2365	0.6938
1.3058	14.9091	123	1.2258	0.6938
1.2181	16.0	132	1.1759	0.6822
1.1537	16.9697	140	1.1413	0.7074
1.1537	17.9394	148	1.0586	0.7248
1.0819	18.9091	156	1.0059	0.7558
0.9905	20.0	165	0.9575	0.7578
1.0055	20.9697	173	0.9807	0.7442
0.9484	21.9394	181	0.9553	0.7539
0.9484	22.9091	189	0.8213	0.8004
0.8974	24.0	198	0.8305	0.8043
0.8545	24.9697	206	0.8273	0.7849
0.8724	25.9394	214	0.8177	0.7519
0.8642	26.9091	222	0.7692	0.7926
0.7609	28.0	231	0.7293	0.8062
0.7609	28.9697	239	0.7001	0.8198
0.7418	29.9394	247	0.7899	0.7636
0.7552	30.9091	255	0.6595	0.8101
0.7291	32.0	264	0.6971	0.7907
0.693	32.9697	272	0.7215	0.7946
0.6891	33.9394	280	0.6980	0.8004
0.6891	34.9091	288	0.6200	0.8372
0.6936	36.0	297	0.7245	0.7733
0.6698	36.9697	305	0.6724	0.7984
0.6502	37.9394	313	0.6701	0.8023
0.6988	38.9091	321	0.6049	0.8236
0.6709	40.0	330	0.6397	0.7965
0.6709	40.9697	338	0.5654	0.8391
0.652	41.9394	346	0.6371	0.8101
0.64	42.9091	354	0.6341	0.8062
0.6368	44.0	363	0.5662	0.8527
0.595	44.9697	371	0.5744	0.8411
0.595	45.9394	379	0.5465	0.8430
0.5823	46.9091	387	0.6254	0.7984
0.5514	48.0	396	0.5368	0.8333
0.5693	48.9697	404	0.5705	0.8043
0.5244	49.9394	412	0.5685	0.8314
0.5495	50.9091	420	0.5811	0.8120
0.5495	52.0	429	0.5037	0.8469
0.5501	52.9697	437	0.5423	0.8372
0.5405	53.9394	445	0.5487	0.8178
0.534	54.9091	453	0.5607	0.8217
0.5502	56.0	462	0.5141	0.8198
0.4772	56.9697	470	0.4813	0.8605
0.4772	57.9394	478	0.5007	0.8566
0.4823	58.9091	486	0.4847	0.8624
0.5107	60.0	495	0.5273	0.8333
0.5205	60.9697	503	0.4981	0.8430
0.5171	61.9394	511	0.4819	0.8430
0.5171	62.9091	519	0.4415	0.8682
0.5498	64.0	528	0.4578	0.8566
0.4732	64.9697	536	0.4614	0.8450
0.4623	65.9394	544	0.4923	0.8488
0.4406	66.9091	552	0.4556	0.8547
0.4889	68.0	561	0.4727	0.8488
0.4889	68.9697	569	0.4746	0.8469
0.4532	69.9394	577	0.4496	0.8585
0.3988	70.9091	585	0.4260	0.8702
0.4608	72.0	594	0.4464	0.8547
0.4429	72.9697	602	0.3946	0.8818
0.4502	73.9394	610	0.4566	0.8527
0.4502	74.9091	618	0.4472	0.8663
0.4381	76.0	627	0.4701	0.8372
0.4437	76.9697	635	0.4351	0.8488
0.4223	77.9394	643	0.4011	0.8779
0.4121	78.9091	651	0.4328	0.8547
0.4164	80.0	660	0.3908	0.8857
0.4164	80.9697	668	0.3774	0.8876
0.418	81.9394	676	0.4397	0.8643
0.3961	82.9091	684	0.4500	0.8585
0.4035	84.0	693	0.3968	0.8624
0.4269	84.9697	701	0.4457	0.8566
0.4269	85.9394	709	0.3987	0.8740
0.3694	86.9091	717	0.4074	0.8760
0.3642	88.0	726	0.3781	0.9012
0.3985	88.9697	734	0.3575	0.8934
0.4237	89.9394	742	0.4313	0.8508
0.4156	90.9091	750	0.3504	0.8934
0.4156	92.0	759	0.4116	0.8566
0.389	92.9697	767	0.3739	0.8779
0.3934	93.9394	775	0.3990	0.8779
0.4231	94.9091	783	0.4164	0.8624
0.3792	96.0	792	0.3808	0.8721
0.3928	96.9697	800	0.3534	0.8915
0.3928	97.9394	808	0.3643	0.8798
0.4003	98.9091	816	0.4150	0.8624
0.3929	100.0	825	0.3477	0.9050
0.3992	100.9697	833	0.4037	0.8682
0.387	101.9394	841	0.3453	0.9050
0.387	102.9091	849	0.4012	0.8682
0.3942	104.0	858	0.3843	0.8915
0.3794	104.9697	866	0.3478	0.8798
0.3794	105.9394	874	0.3111	0.9167
0.396	106.9091	882	0.3588	0.8818
0.3767	108.0	891	0.3602	0.8837
0.3767	108.6061	896	0.3784	0.8760

Framework versions

• Transformers 4.45.0.dev0
• Pytorch 2.4.0+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1