deit-base-distilled-patch16-224-hasta-75-fold5

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: 0.6803
• 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	1.1508	0.4167
No log	2.0	2	0.9687	0.4167
No log	3.0	3	0.6803	0.9167
No log	4.0	4	0.4213	0.9167
No log	5.0	5	0.2979	0.9167
No log	6.0	6	0.2848	0.9167
No log	7.0	7	0.3062	0.9167
No log	8.0	8	0.3740	0.9167
No log	9.0	9	0.4779	0.9167
0.3289	10.0	10	0.4269	0.9167
0.3289	11.0	11	0.2878	0.9167
0.3289	12.0	12	0.2469	0.9167
0.3289	13.0	13	0.2483	0.9167
0.3289	14.0	14	0.2873	0.9167
0.3289	15.0	15	0.3418	0.9167
0.3289	16.0	16	0.3439	0.9167
0.3289	17.0	17	0.3146	0.9167
0.3289	18.0	18	0.3273	0.9167
0.3289	19.0	19	0.3637	0.9167
0.1373	20.0	20	0.4029	0.9167
0.1373	21.0	21	0.3972	0.9167
0.1373	22.0	22	0.3716	0.9167
0.1373	23.0	23	0.3374	0.9167
0.1373	24.0	24	0.2715	0.9167
0.1373	25.0	25	0.2576	0.9167
0.1373	26.0	26	0.2621	0.9167
0.1373	27.0	27	0.2693	0.9167
0.1373	28.0	28	0.2653	0.9167
0.1373	29.0	29	0.2811	0.8333
0.0748	30.0	30	0.3335	0.8333
0.0748	31.0	31	0.3538	0.8333
0.0748	32.0	32	0.2896	0.8333
0.0748	33.0	33	0.2242	0.9167
0.0748	34.0	34	0.2258	0.9167
0.0748	35.0	35	0.2552	0.9167
0.0748	36.0	36	0.3066	0.8333
0.0748	37.0	37	0.3862	0.8333
0.0748	38.0	38	0.4926	0.8333
0.0748	39.0	39	0.6405	0.75
0.0581	40.0	40	0.6536	0.75
0.0581	41.0	41	0.6079	0.8333
0.0581	42.0	42	0.4895	0.8333
0.0581	43.0	43	0.4001	0.8333
0.0581	44.0	44	0.3493	0.8333
0.0581	45.0	45	0.3068	0.8333
0.0581	46.0	46	0.2859	0.9167
0.0581	47.0	47	0.2943	0.9167
0.0581	48.0	48	0.2983	0.9167
0.0581	49.0	49	0.3183	0.9167
0.041	50.0	50	0.3299	0.9167
0.041	51.0	51	0.3353	0.9167
0.041	52.0	52	0.3243	0.9167
0.041	53.0	53	0.3057	0.9167
0.041	54.0	54	0.2817	0.9167
0.041	55.0	55	0.2535	0.9167
0.041	56.0	56	0.2444	0.9167
0.041	57.0	57	0.2356	0.9167
0.041	58.0	58	0.2455	0.9167
0.041	59.0	59	0.2519	0.9167
0.0333	60.0	60	0.2529	0.9167
0.0333	61.0	61	0.2582	0.9167
0.0333	62.0	62	0.2589	0.9167
0.0333	63.0	63	0.2585	0.9167
0.0333	64.0	64	0.2540	0.9167
0.0333	65.0	65	0.2631	0.9167
0.0333	66.0	66	0.2697	0.9167
0.0333	67.0	67	0.2714	0.9167
0.0333	68.0	68	0.2740	0.9167
0.0333	69.0	69	0.2738	0.9167
0.0221	70.0	70	0.2803	0.9167
0.0221	71.0	71	0.2755	0.9167
0.0221	72.0	72	0.2711	0.9167
0.0221	73.0	73	0.2706	0.9167
0.0221	74.0	74	0.2752	0.9167
0.0221	75.0	75	0.2650	0.9167
0.0221	76.0	76	0.2605	0.9167
0.0221	77.0	77	0.2650	0.9167
0.0221	78.0	78	0.2731	0.9167
0.0221	79.0	79	0.2746	0.9167
0.0151	80.0	80	0.2776	0.9167
0.0151	81.0	81	0.2773	0.9167
0.0151	82.0	82	0.2793	0.9167
0.0151	83.0	83	0.2747	0.9167
0.0151	84.0	84	0.2755	0.9167
0.0151	85.0	85	0.2746	0.9167
0.0151	86.0	86	0.2741	0.9167
0.0151	87.0	87	0.2739	0.9167
0.0151	88.0	88	0.2781	0.9167
0.0151	89.0	89	0.2838	0.9167
0.0213	90.0	90	0.2855	0.9167
0.0213	91.0	91	0.2862	0.9167
0.0213	92.0	92	0.2857	0.9167
0.0213	93.0	93	0.2824	0.9167
0.0213	94.0	94	0.2758	0.9167
0.0213	95.0	95	0.2698	0.9167
0.0213	96.0	96	0.2625	0.9167
0.0213	97.0	97	0.2566	0.9167
0.0213	98.0	98	0.2517	0.9167
0.0213	99.0	99	0.2491	0.9167
0.0314	100.0	100	0.2478	0.9167

Framework versions

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