beit-base-patch16-224-85-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: 0.2763
• Accuracy: 0.9318

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	0.6057	0.7273
No log	2.0	4	0.6639	0.7045
No log	3.0	6	0.7324	0.7045
No log	4.0	8	0.5213	0.7273
0.5701	5.0	10	0.4717	0.8182
0.5701	6.0	12	0.5339	0.7045
0.5701	7.0	14	0.4959	0.7273
0.5701	8.0	16	0.4086	0.8409
0.5701	9.0	18	0.4039	0.8182
0.4248	10.0	20	0.4106	0.8182
0.4248	11.0	22	0.4108	0.8409
0.4248	12.0	24	0.4607	0.7727
0.4248	13.0	26	0.4446	0.7727
0.4248	14.0	28	0.3912	0.8409
0.3579	15.0	30	0.5183	0.7727
0.3579	16.0	32	0.2991	0.8864
0.3579	17.0	34	0.3587	0.8182
0.3579	18.0	36	0.3110	0.8182
0.3579	19.0	38	0.3084	0.8636
0.2838	20.0	40	0.3079	0.8864
0.2838	21.0	42	0.3033	0.8409
0.2838	22.0	44	0.3126	0.8409
0.2838	23.0	46	0.3171	0.8864
0.2838	24.0	48	0.2689	0.8636
0.2705	25.0	50	0.3175	0.8409
0.2705	26.0	52	0.3464	0.8409
0.2705	27.0	54	0.3092	0.8636
0.2705	28.0	56	0.3178	0.8636
0.2705	29.0	58	0.4107	0.7955
0.1887	30.0	60	0.4151	0.8182
0.1887	31.0	62	0.5450	0.7955
0.1887	32.0	64	0.2892	0.8409
0.1887	33.0	66	0.4078	0.8409
0.1887	34.0	68	0.2821	0.8636
0.1692	35.0	70	0.2708	0.8636
0.1692	36.0	72	0.2692	0.8864
0.1692	37.0	74	0.2806	0.8864
0.1692	38.0	76	0.4613	0.8182
0.1692	39.0	78	0.2887	0.9091
0.1623	40.0	80	0.4046	0.8409
0.1623	41.0	82	0.4542	0.8409
0.1623	42.0	84	0.3010	0.8636
0.1623	43.0	86	0.2954	0.8636
0.1623	44.0	88	0.2838	0.8864
0.1522	45.0	90	0.2675	0.8864
0.1522	46.0	92	0.2517	0.9091
0.1522	47.0	94	0.2687	0.9091
0.1522	48.0	96	0.2551	0.9091
0.1522	49.0	98	0.2661	0.8864
0.1379	50.0	100	0.3507	0.8182
0.1379	51.0	102	0.2629	0.8864
0.1379	52.0	104	0.2697	0.8864
0.1379	53.0	106	0.3081	0.8636
0.1379	54.0	108	0.3851	0.8409
0.1283	55.0	110	0.3104	0.8636
0.1283	56.0	112	0.3624	0.8864
0.1283	57.0	114	0.3199	0.8864
0.1283	58.0	116	0.4964	0.8182
0.1283	59.0	118	0.3356	0.8864
0.1335	60.0	120	0.2314	0.9091
0.1335	61.0	122	0.2334	0.9091
0.1335	62.0	124	0.3961	0.8636
0.1335	63.0	126	0.3453	0.8636
0.1335	64.0	128	0.2806	0.8636
0.1353	65.0	130	0.3372	0.8636
0.1353	66.0	132	0.2675	0.8864
0.1353	67.0	134	0.3482	0.8864
0.1353	68.0	136	0.3725	0.8636
0.1353	69.0	138	0.3769	0.8636
0.099	70.0	140	0.5170	0.8409
0.099	71.0	142	0.4710	0.8636
0.099	72.0	144	0.3266	0.9091
0.099	73.0	146	0.3390	0.8636
0.099	74.0	148	0.3051	0.8636
0.1179	75.0	150	0.3030	0.9091
0.1179	76.0	152	0.3208	0.9091
0.1179	77.0	154	0.2954	0.9091
0.1179	78.0	156	0.2777	0.9091
0.1179	79.0	158	0.2763	0.9318
0.1077	80.0	160	0.3059	0.9091
0.1077	81.0	162	0.3445	0.8864
0.1077	82.0	164	0.3239	0.9091
0.1077	83.0	166	0.3175	0.9091
0.1077	84.0	168	0.3214	0.9091
0.0907	85.0	170	0.3313	0.9091
0.0907	86.0	172	0.3492	0.9091
0.0907	87.0	174	0.3644	0.9091
0.0907	88.0	176	0.3637	0.9091
0.0907	89.0	178	0.3750	0.9091
0.0972	90.0	180	0.3845	0.9091
0.0972	91.0	182	0.3749	0.9091
0.0972	92.0	184	0.3721	0.8864
0.0972	93.0	186	0.3680	0.8864
0.0972	94.0	188	0.3634	0.8864
0.0733	95.0	190	0.3565	0.9091
0.0733	96.0	192	0.3519	0.9091
0.0733	97.0	194	0.3529	0.9091
0.0733	98.0	196	0.3536	0.9091
0.0733	99.0	198	0.3561	0.9091
0.079	100.0	200	0.3565	0.9091

Framework versions

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