deit-base-distilled-patch16-224-hasta-75-fold2

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

• Loss: 0.5330
• Accuracy: 0.9167

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	1	0.8721	0.5833
No log	2.0	2	0.7324	0.8333
No log	3.0	3	0.5330	0.9167
No log	4.0	4	0.4022	0.9167
No log	5.0	5	0.3794	0.9167
No log	6.0	6	0.3910	0.9167
No log	7.0	7	0.3832	0.9167
No log	8.0	8	0.3524	0.9167
No log	9.0	9	0.4280	0.9167
0.3237	10.0	10	0.5286	0.8333
0.3237	11.0	11	0.4004	0.9167
0.3237	12.0	12	0.3327	0.9167
0.3237	13.0	13	0.3136	0.9167
0.3237	14.0	14	0.2844	0.9167
0.3237	15.0	15	0.2493	0.9167
0.3237	16.0	16	0.2190	0.9167
0.3237	17.0	17	0.1925	0.9167
0.3237	18.0	18	0.1722	0.9167
0.3237	19.0	19	0.1387	0.9167
0.142	20.0	20	0.1259	0.9167
0.142	21.0	21	0.1443	0.9167
0.142	22.0	22	0.1372	0.9167
0.142	23.0	23	0.1043	0.9167
0.142	24.0	24	0.1022	0.9167
0.142	25.0	25	0.1327	0.9167
0.142	26.0	26	0.2213	0.9167
0.142	27.0	27	0.2587	0.9167
0.142	28.0	28	0.2411	0.9167
0.142	29.0	29	0.1915	0.9167
0.0723	30.0	30	0.1418	0.9167
0.0723	31.0	31	0.1369	0.9167
0.0723	32.0	32	0.1749	0.9167
0.0723	33.0	33	0.2607	0.9167
0.0723	34.0	34	0.3049	0.9167
0.0723	35.0	35	0.3103	0.9167
0.0723	36.0	36	0.2972	0.9167
0.0723	37.0	37	0.2901	0.9167
0.0723	38.0	38	0.2490	0.9167
0.0723	39.0	39	0.2047	0.9167
0.0458	40.0	40	0.1781	0.9167
0.0458	41.0	41	0.1712	0.9167
0.0458	42.0	42	0.2114	0.9167
0.0458	43.0	43	0.2837	0.9167
0.0458	44.0	44	0.3335	0.9167
0.0458	45.0	45	0.3600	0.9167
0.0458	46.0	46	0.3698	0.9167
0.0458	47.0	47	0.3607	0.9167
0.0458	48.0	48	0.3493	0.9167
0.0458	49.0	49	0.3408	0.9167
0.0478	50.0	50	0.3538	0.9167
0.0478	51.0	51	0.3481	0.9167
0.0478	52.0	52	0.3513	0.9167
0.0478	53.0	53	0.3336	0.9167
0.0478	54.0	54	0.3044	0.9167
0.0478	55.0	55	0.2844	0.9167
0.0478	56.0	56	0.2790	0.9167
0.0478	57.0	57	0.2990	0.9167
0.0478	58.0	58	0.3265	0.9167
0.0478	59.0	59	0.3682	0.9167
0.0145	60.0	60	0.3938	0.9167
0.0145	61.0	61	0.4018	0.9167
0.0145	62.0	62	0.3817	0.9167
0.0145	63.0	63	0.3376	0.9167
0.0145	64.0	64	0.2812	0.9167
0.0145	65.0	65	0.2029	0.9167
0.0145	66.0	66	0.1343	0.9167
0.0145	67.0	67	0.0996	0.9167
0.0145	68.0	68	0.0811	0.9167
0.0145	69.0	69	0.0662	0.9167
0.0447	70.0	70	0.0745	0.9167
0.0447	71.0	71	0.1053	0.9167
0.0447	72.0	72	0.1643	0.9167
0.0447	73.0	73	0.2353	0.9167
0.0447	74.0	74	0.3155	0.9167
0.0447	75.0	75	0.3678	0.9167
0.0447	76.0	76	0.3946	0.9167
0.0447	77.0	77	0.4025	0.9167
0.0447	78.0	78	0.4106	0.9167
0.0447	79.0	79	0.4147	0.9167
0.0229	80.0	80	0.4108	0.9167
0.0229	81.0	81	0.3993	0.9167
0.0229	82.0	82	0.3857	0.9167
0.0229	83.0	83	0.3644	0.9167
0.0229	84.0	84	0.3422	0.9167
0.0229	85.0	85	0.3280	0.9167
0.0229	86.0	86	0.3108	0.9167
0.0229	87.0	87	0.2936	0.9167
0.0229	88.0	88	0.2846	0.9167
0.0229	89.0	89	0.2861	0.9167
0.0317	90.0	90	0.2909	0.9167
0.0317	91.0	91	0.2921	0.9167
0.0317	92.0	92	0.2938	0.9167
0.0317	93.0	93	0.2974	0.9167
0.0317	94.0	94	0.2998	0.9167
0.0317	95.0	95	0.2994	0.9167
0.0317	96.0	96	0.2992	0.9167
0.0317	97.0	97	0.2973	0.9167
0.0317	98.0	98	0.2969	0.9167
0.0317	99.0	99	0.2970	0.9167
0.0283	100.0	100	0.2973	0.9167

Framework versions

• Transformers 4.41.0
• Pytorch 2.3.0+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1