beit-base-patch16-224-75-fold2

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

• Loss: 0.2685
• Accuracy: 0.9535

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	1.0	2	0.7091	0.5349
No log	2.0	4	0.6502	0.7209
No log	3.0	6	0.9193	0.6977
No log	4.0	8	0.7499	0.7442
0.6436	5.0	10	0.4527	0.8140
0.6436	6.0	12	0.4169	0.8372
0.6436	7.0	14	0.5773	0.7442
0.6436	8.0	16	0.4076	0.8605
0.6436	9.0	18	0.3939	0.8605
0.3863	10.0	20	0.4017	0.8605
0.3863	11.0	22	0.4918	0.8140
0.3863	12.0	24	0.2688	0.8372
0.3863	13.0	26	0.3884	0.8140
0.3863	14.0	28	0.3679	0.8140
0.2925	15.0	30	0.2802	0.8837
0.2925	16.0	32	0.2436	0.9070
0.2925	17.0	34	0.2337	0.9302
0.2925	18.0	36	0.3711	0.8140
0.2925	19.0	38	0.2372	0.9302
0.2289	20.0	40	0.2685	0.9535
0.2289	21.0	42	0.2610	0.9070
0.2289	22.0	44	0.3328	0.8372
0.2289	23.0	46	0.3479	0.8372
0.2289	24.0	48	0.2855	0.8837
0.219	25.0	50	0.2962	0.9070
0.219	26.0	52	0.4038	0.9070
0.219	27.0	54	0.3149	0.9070
0.219	28.0	56	0.3212	0.9070
0.219	29.0	58	0.4895	0.8605
0.1933	30.0	60	0.4335	0.8837
0.1933	31.0	62	0.3521	0.8372
0.1933	32.0	64	0.2960	0.8837
0.1933	33.0	66	0.4037	0.8372
0.1933	34.0	68	0.2913	0.8837
0.1892	35.0	70	0.3043	0.8837
0.1892	36.0	72	0.3602	0.9302
0.1892	37.0	74	0.3315	0.9302
0.1892	38.0	76	0.2674	0.9302
0.1892	39.0	78	0.2970	0.9535
0.15	40.0	80	0.2661	0.9535
0.15	41.0	82	0.2551	0.8837
0.15	42.0	84	0.2467	0.9302
0.15	43.0	86	0.3008	0.9535
0.15	44.0	88	0.3265	0.9302
0.1238	45.0	90	0.2668	0.9302
0.1238	46.0	92	0.2574	0.9302
0.1238	47.0	94	0.2498	0.9535
0.1238	48.0	96	0.3319	0.8837
0.1238	49.0	98	0.2358	0.9302
0.1063	50.0	100	0.2015	0.9302
0.1063	51.0	102	0.2171	0.9302
0.1063	52.0	104	0.3119	0.9302
0.1063	53.0	106	0.2674	0.9070
0.1063	54.0	108	0.3076	0.8837
0.1112	55.0	110	0.3182	0.8837
0.1112	56.0	112	0.3371	0.9070
0.1112	57.0	114	0.3540	0.9070
0.1112	58.0	116	0.4058	0.9070
0.1112	59.0	118	0.4013	0.9070
0.1128	60.0	120	0.3309	0.9302
0.1128	61.0	122	0.3272	0.9302
0.1128	62.0	124	0.4012	0.9070
0.1128	63.0	126	0.5794	0.8605
0.1128	64.0	128	0.3881	0.9070
0.1168	65.0	130	0.2990	0.9070
0.1168	66.0	132	0.3018	0.8837
0.1168	67.0	134	0.2561	0.9302
0.1168	68.0	136	0.2921	0.9302
0.1168	69.0	138	0.3258	0.9070
0.0846	70.0	140	0.2925	0.9302
0.0846	71.0	142	0.3073	0.9302
0.0846	72.0	144	0.3318	0.9302
0.0846	73.0	146	0.3427	0.9302
0.0846	74.0	148	0.3588	0.9070
0.0845	75.0	150	0.3939	0.9070
0.0845	76.0	152	0.3774	0.9070
0.0845	77.0	154	0.3746	0.9070
0.0845	78.0	156	0.4073	0.8837
0.0845	79.0	158	0.3886	0.9070
0.0885	80.0	160	0.3765	0.9070
0.0885	81.0	162	0.3977	0.9070
0.0885	82.0	164	0.3864	0.9070
0.0885	83.0	166	0.3809	0.9070
0.0885	84.0	168	0.4492	0.8605
0.0859	85.0	170	0.5479	0.8605
0.0859	86.0	172	0.5372	0.8605
0.0859	87.0	174	0.4512	0.8605
0.0859	88.0	176	0.3930	0.9070
0.0859	89.0	178	0.3842	0.9302
0.0764	90.0	180	0.3808	0.9302
0.0764	91.0	182	0.3787	0.9302
0.0764	92.0	184	0.3833	0.9070
0.0764	93.0	186	0.3912	0.9070
0.0764	94.0	188	0.3888	0.8837
0.0727	95.0	190	0.3817	0.8837
0.0727	96.0	192	0.3708	0.9070
0.0727	97.0	194	0.3640	0.9070
0.0727	98.0	196	0.3613	0.9302
0.0727	99.0	198	0.3607	0.9302
0.069	100.0	200	0.3605	0.9302

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1