beit-base-patch16-224-75-fold1

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: 0.2641
• Accuracy: 0.9302

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	1.0987	0.3023
No log	2.0	4	0.6630	0.6977
No log	3.0	6	0.8342	0.6977
No log	4.0	8	0.6752	0.6977
0.7768	5.0	10	0.5408	0.7209
0.7768	6.0	12	0.7252	0.6977
0.7768	7.0	14	0.5609	0.7209
0.7768	8.0	16	0.7345	0.6977
0.7768	9.0	18	0.4614	0.7674
0.4249	10.0	20	0.4434	0.8372
0.4249	11.0	22	0.7552	0.7442
0.4249	12.0	24	0.4142	0.7674
0.4249	13.0	26	0.7183	0.7442
0.4249	14.0	28	0.5591	0.7907
0.3506	15.0	30	0.4363	0.6977
0.3506	16.0	32	0.5738	0.7907
0.3506	17.0	34	0.4286	0.8140
0.3506	18.0	36	0.4200	0.8140
0.3506	19.0	38	0.6514	0.7442
0.3434	20.0	40	0.4190	0.7907
0.3434	21.0	42	0.6220	0.8140
0.3434	22.0	44	0.6334	0.7907
0.3434	23.0	46	0.4487	0.8372
0.3434	24.0	48	0.4960	0.8605
0.2498	25.0	50	0.4179	0.8605
0.2498	26.0	52	0.3221	0.8605
0.2498	27.0	54	0.4776	0.8372
0.2498	28.0	56	0.5756	0.8605
0.2498	29.0	58	0.5444	0.8372
0.2461	30.0	60	0.3973	0.8605
0.2461	31.0	62	0.3672	0.8605
0.2461	32.0	64	0.4071	0.8837
0.2461	33.0	66	0.4678	0.7674
0.2461	34.0	68	0.2641	0.9302
0.2279	35.0	70	0.5551	0.8372
0.2279	36.0	72	0.2727	0.9302
0.2279	37.0	74	0.3312	0.8837
0.2279	38.0	76	0.7485	0.7907
0.2279	39.0	78	0.6407	0.8605
0.183	40.0	80	0.5420	0.8372
0.183	41.0	82	0.7364	0.8605
0.183	42.0	84	0.4141	0.8605
0.183	43.0	86	0.5461	0.7907
0.183	44.0	88	0.3438	0.8605
0.1658	45.0	90	0.3322	0.9302
0.1658	46.0	92	0.3463	0.9302
0.1658	47.0	94	0.6066	0.8605
0.1658	48.0	96	0.6259	0.8605
0.1658	49.0	98	0.4909	0.8372
0.1555	50.0	100	0.6022	0.7907
0.1555	51.0	102	0.5234	0.8372
0.1555	52.0	104	0.4164	0.8837
0.1555	53.0	106	0.3893	0.8605
0.1555	54.0	108	0.3774	0.8837
0.1487	55.0	110	0.7532	0.8372
0.1487	56.0	112	0.7141	0.8605
0.1487	57.0	114	0.4197	0.9070
0.1487	58.0	116	0.6816	0.7442
0.1487	59.0	118	0.5384	0.8140
0.1349	60.0	120	0.4971	0.8605
0.1349	61.0	122	0.4601	0.8837
0.1349	62.0	124	0.4740	0.8372
0.1349	63.0	126	0.5386	0.8140
0.1349	64.0	128	0.3376	0.9070
0.128	65.0	130	0.3905	0.9070
0.128	66.0	132	0.3841	0.9302
0.128	67.0	134	0.3567	0.8605
0.128	68.0	136	0.3985	0.8372
0.128	69.0	138	0.4165	0.8372
0.0875	70.0	140	0.4346	0.8605
0.0875	71.0	142	0.4497	0.8372
0.0875	72.0	144	0.4353	0.8837
0.0875	73.0	146	0.4276	0.8837
0.0875	74.0	148	0.4010	0.8837
0.0932	75.0	150	0.3958	0.9070
0.0932	76.0	152	0.3604	0.9070
0.0932	77.0	154	0.3427	0.8837
0.0932	78.0	156	0.3417	0.8837
0.0932	79.0	158	0.3438	0.9070
0.0943	80.0	160	0.3756	0.9302
0.0943	81.0	162	0.4077	0.9302
0.0943	82.0	164	0.4129	0.9302
0.0943	83.0	166	0.4304	0.9302
0.0943	84.0	168	0.4156	0.9302
0.0753	85.0	170	0.4088	0.9070
0.0753	86.0	172	0.4090	0.8837
0.0753	87.0	174	0.4076	0.9070
0.0753	88.0	176	0.4273	0.9070
0.0753	89.0	178	0.4367	0.9070
0.0846	90.0	180	0.4490	0.9070
0.0846	91.0	182	0.4448	0.8837
0.0846	92.0	184	0.4406	0.8837
0.0846	93.0	186	0.4393	0.8837
0.0846	94.0	188	0.4370	0.8837
0.0865	95.0	190	0.4330	0.8837
0.0865	96.0	192	0.4293	0.8837
0.0865	97.0	194	0.4240	0.8837
0.0865	98.0	196	0.4177	0.8837
0.0865	99.0	198	0.4144	0.8837
0.1019	100.0	200	0.4135	0.8837

Framework versions

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