beit-base-patch16-224-75-fold4

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: 0.2509
• 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	0.5130	0.7907
No log	2.0	4	0.4861	0.7907
No log	3.0	6	0.4775	0.7907
No log	4.0	8	0.4419	0.7907
0.4909	5.0	10	0.3672	0.8605
0.4909	6.0	12	0.3301	0.8837
0.4909	7.0	14	0.3131	0.8837
0.4909	8.0	16	0.4535	0.8605
0.4909	9.0	18	0.3088	0.8372
0.3473	10.0	20	0.4453	0.8837
0.3473	11.0	22	0.4234	0.8605
0.3473	12.0	24	0.3601	0.8837
0.3473	13.0	26	0.3658	0.9070
0.3473	14.0	28	0.3081	0.8837
0.2903	15.0	30	0.4128	0.8837
0.2903	16.0	32	0.2555	0.8605
0.2903	17.0	34	0.3341	0.8837
0.2903	18.0	36	0.2427	0.8837
0.2903	19.0	38	0.4325	0.8372
0.2673	20.0	40	0.2637	0.9070
0.2673	21.0	42	0.2919	0.8837
0.2673	22.0	44	0.3139	0.8837
0.2673	23.0	46	0.2411	0.8837
0.2673	24.0	48	0.4645	0.9070
0.2103	25.0	50	0.5084	0.8605
0.2103	26.0	52	0.2308	0.9070
0.2103	27.0	54	0.3450	0.8605
0.2103	28.0	56	0.3444	0.8605
0.2103	29.0	58	0.2546	0.9070
0.1673	30.0	60	0.9117	0.8140
0.1673	31.0	62	0.8437	0.8140
0.1673	32.0	64	0.6758	0.8372
0.1673	33.0	66	0.8019	0.8140
0.1673	34.0	68	0.3364	0.8837
0.1677	35.0	70	0.2928	0.8837
0.1677	36.0	72	0.2547	0.9070
0.1677	37.0	74	0.2969	0.8837
0.1677	38.0	76	0.5706	0.8837
0.1677	39.0	78	0.7006	0.8837
0.1407	40.0	80	0.4321	0.8837
0.1407	41.0	82	0.4366	0.8837
0.1407	42.0	84	0.3956	0.8837
0.1407	43.0	86	0.2290	0.8372
0.1407	44.0	88	0.3665	0.8837
0.1474	45.0	90	0.4465	0.8605
0.1474	46.0	92	0.7279	0.8605
0.1474	47.0	94	0.5259	0.8605
0.1474	48.0	96	0.5832	0.8837
0.1474	49.0	98	0.7328	0.8837
0.1344	50.0	100	0.3890	0.8837
0.1344	51.0	102	0.2642	0.8837
0.1344	52.0	104	0.3710	0.9070
0.1344	53.0	106	0.4773	0.9070
0.1344	54.0	108	0.3628	0.9302
0.1166	55.0	110	0.4389	0.9070
0.1166	56.0	112	0.4813	0.9070
0.1166	57.0	114	0.5328	0.9070
0.1166	58.0	116	0.5342	0.9070
0.1166	59.0	118	0.4892	0.9070
0.097	60.0	120	0.5857	0.9070
0.097	61.0	122	0.6681	0.9070
0.097	62.0	124	0.5947	0.9070
0.097	63.0	126	0.4749	0.9070
0.097	64.0	128	0.6091	0.8837
0.1076	65.0	130	0.9725	0.8605
0.1076	66.0	132	1.1372	0.8140
0.1076	67.0	134	0.7109	0.8605
0.1076	68.0	136	0.3549	0.9302
0.1076	69.0	138	0.2709	0.9302
0.0914	70.0	140	0.3316	0.9302
0.0914	71.0	142	0.3176	0.9302
0.0914	72.0	144	0.2509	0.9535
0.0914	73.0	146	0.2256	0.9070
0.0914	74.0	148	0.2570	0.9070
0.0815	75.0	150	0.3081	0.9535
0.0815	76.0	152	0.4199	0.9302
0.0815	77.0	154	0.4324	0.9302
0.0815	78.0	156	0.3928	0.9302
0.0815	79.0	158	0.3700	0.9302
0.0878	80.0	160	0.3812	0.9302
0.0878	81.0	162	0.4300	0.9302
0.0878	82.0	164	0.4289	0.9302
0.0878	83.0	166	0.4125	0.9302
0.0878	84.0	168	0.4351	0.9302
0.0725	85.0	170	0.5046	0.9302
0.0725	86.0	172	0.5692	0.9070
0.0725	87.0	174	0.5486	0.9070
0.0725	88.0	176	0.5310	0.9302
0.0725	89.0	178	0.4662	0.9302
0.0944	90.0	180	0.4070	0.9302
0.0944	91.0	182	0.3768	0.9302
0.0944	92.0	184	0.3884	0.9302
0.0944	93.0	186	0.3851	0.9302
0.0944	94.0	188	0.3759	0.9302
0.0739	95.0	190	0.3608	0.9302
0.0739	96.0	192	0.3456	0.9302
0.0739	97.0	194	0.3360	0.9302
0.0739	98.0	196	0.3312	0.9302
0.0739	99.0	198	0.3321	0.9302
0.0612	100.0	200	0.3331	0.9302

Framework versions

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