deit-base-distilled-patch16-224-hasta-75-fold4

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: 0.4466
• 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	1.3783	0.0
No log	2.0	2	1.1470	0.1667
No log	3.0	3	0.7684	0.75
No log	4.0	4	0.4466	0.9167
No log	5.0	5	0.3116	0.9167
No log	6.0	6	0.3024	0.9167
No log	7.0	7	0.3113	0.9167
No log	8.0	8	0.3526	0.9167
No log	9.0	9	0.5370	0.9167
0.3843	10.0	10	0.6259	0.8333
0.3843	11.0	11	0.4979	0.9167
0.3843	12.0	12	0.3555	0.9167
0.3843	13.0	13	0.3132	0.9167
0.3843	14.0	14	0.3054	0.9167
0.3843	15.0	15	0.3262	0.9167
0.3843	16.0	16	0.3495	0.9167
0.3843	17.0	17	0.3211	0.9167
0.3843	18.0	18	0.2993	0.9167
0.3843	19.0	19	0.3174	0.9167
0.1412	20.0	20	0.3221	0.8333
0.1412	21.0	21	0.3248	0.8333
0.1412	22.0	22	0.3245	0.8333
0.1412	23.0	23	0.3412	0.8333
0.1412	24.0	24	0.3021	0.8333
0.1412	25.0	25	0.2038	0.9167
0.1412	26.0	26	0.1856	0.9167
0.1412	27.0	27	0.2126	0.9167
0.1412	28.0	28	0.2161	0.9167
0.1412	29.0	29	0.1838	0.9167
0.0596	30.0	30	0.1688	0.9167
0.0596	31.0	31	0.1827	0.9167
0.0596	32.0	32	0.1860	0.9167
0.0596	33.0	33	0.1819	0.9167
0.0596	34.0	34	0.1868	0.9167
0.0596	35.0	35	0.2212	0.8333
0.0596	36.0	36	0.2478	0.8333
0.0596	37.0	37	0.2653	0.8333
0.0596	38.0	38	0.2093	0.9167
0.0596	39.0	39	0.1924	0.9167
0.0541	40.0	40	0.1789	0.9167
0.0541	41.0	41	0.1646	0.9167
0.0541	42.0	42	0.1635	0.9167
0.0541	43.0	43	0.1611	0.9167
0.0541	44.0	44	0.1592	0.9167
0.0541	45.0	45	0.1754	0.9167
0.0541	46.0	46	0.1908	0.9167
0.0541	47.0	47	0.1859	0.9167
0.0541	48.0	48	0.1687	0.9167
0.0541	49.0	49	0.1646	0.9167
0.0306	50.0	50	0.1663	0.9167
0.0306	51.0	51	0.1609	0.9167
0.0306	52.0	52	0.1791	0.9167
0.0306	53.0	53	0.2029	0.9167
0.0306	54.0	54	0.2205	0.9167
0.0306	55.0	55	0.2358	0.9167
0.0306	56.0	56	0.2392	0.9167
0.0306	57.0	57	0.2591	0.9167
0.0306	58.0	58	0.2536	0.9167
0.0306	59.0	59	0.2678	0.9167
0.0369	60.0	60	0.2655	0.9167
0.0369	61.0	61	0.2782	0.9167
0.0369	62.0	62	0.3050	0.9167
0.0369	63.0	63	0.3199	0.9167
0.0369	64.0	64	0.3130	0.9167
0.0369	65.0	65	0.3063	0.9167
0.0369	66.0	66	0.2885	0.9167
0.0369	67.0	67	0.2654	0.9167
0.0369	68.0	68	0.2478	0.9167
0.0369	69.0	69	0.2358	0.9167
0.0241	70.0	70	0.2106	0.9167
0.0241	71.0	71	0.2047	0.9167
0.0241	72.0	72	0.2100	0.9167
0.0241	73.0	73	0.2092	0.9167
0.0241	74.0	74	0.2261	0.9167
0.0241	75.0	75	0.2380	0.9167
0.0241	76.0	76	0.2644	0.9167
0.0241	77.0	77	0.2972	0.9167
0.0241	78.0	78	0.3053	0.9167
0.0241	79.0	79	0.3133	0.9167
0.0234	80.0	80	0.3051	0.9167
0.0234	81.0	81	0.3001	0.9167
0.0234	82.0	82	0.2921	0.9167
0.0234	83.0	83	0.2899	0.9167
0.0234	84.0	84	0.2798	0.9167
0.0234	85.0	85	0.2641	0.9167
0.0234	86.0	86	0.2514	0.9167
0.0234	87.0	87	0.2419	0.9167
0.0234	88.0	88	0.2282	0.9167
0.0234	89.0	89	0.2174	0.9167
0.0197	90.0	90	0.2070	0.9167
0.0197	91.0	91	0.2006	0.9167
0.0197	92.0	92	0.1977	0.9167
0.0197	93.0	93	0.1956	0.9167
0.0197	94.0	94	0.1946	0.9167
0.0197	95.0	95	0.1931	0.9167
0.0197	96.0	96	0.1920	0.9167
0.0197	97.0	97	0.1922	0.9167
0.0197	98.0	98	0.1926	0.9167
0.0197	99.0	99	0.1942	0.9167
0.0295	100.0	100	0.1950	0.9167

Framework versions

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