Fer_vit_jaffe_GOOGLE_0

This model is a fine-tuned version of WinKawaks/vit-tiny-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.4000
• Accuracy: 0.8667

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: 0
• 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	2.2045	0.1
No log	2.0	2	2.0846	0.1333
No log	3.0	3	2.1060	0.1
No log	4.0	4	1.9913	0.1333
No log	5.0	5	1.9980	0.1
No log	6.0	6	1.8330	0.3
No log	7.0	7	1.9518	0.1333
No log	8.0	8	1.9296	0.1667
No log	9.0	9	1.8688	0.3333
1.9932	10.0	10	1.7509	0.3667
1.9932	11.0	11	1.6357	0.4
1.9932	12.0	12	1.5627	0.3667
1.9932	13.0	13	1.6459	0.3
1.9932	14.0	14	1.5215	0.4
1.9932	15.0	15	1.5421	0.3667
1.9932	16.0	16	1.4164	0.5
1.9932	17.0	17	1.4463	0.3667
1.9932	18.0	18	1.2905	0.4667
1.9932	19.0	19	1.2456	0.6
1.2161	20.0	20	1.2170	0.5667
1.2161	21.0	21	1.0307	0.6
1.2161	22.0	22	1.1198	0.6
1.2161	23.0	23	1.1648	0.5
1.2161	24.0	24	1.0260	0.6
1.2161	25.0	25	1.3020	0.5
1.2161	26.0	26	0.9796	0.6333
1.2161	27.0	27	0.9824	0.6667
1.2161	28.0	28	0.8884	0.7
1.2161	29.0	29	0.9246	0.6333
0.5116	30.0	30	0.8455	0.7333
0.5116	31.0	31	0.7960	0.7
0.5116	32.0	32	0.8179	0.7333
0.5116	33.0	33	0.8721	0.6667
0.5116	34.0	34	0.8279	0.7667
0.5116	35.0	35	0.6486	0.7667
0.5116	36.0	36	0.6816	0.7333
0.5116	37.0	37	0.8016	0.7333
0.5116	38.0	38	0.6464	0.8
0.5116	39.0	39	0.6922	0.7667
0.2101	40.0	40	0.6768	0.7667
0.2101	41.0	41	0.6408	0.7667
0.2101	42.0	42	0.5335	0.8333
0.2101	43.0	43	0.4862	0.8333
0.2101	44.0	44	0.3713	0.8667
0.2101	45.0	45	0.4382	0.8333
0.2101	46.0	46	0.6664	0.7667
0.2101	47.0	47	0.4865	0.8333
0.2101	48.0	48	0.4411	0.8
0.2101	49.0	49	0.4707	0.8667
0.0921	50.0	50	0.6355	0.7667
0.0921	51.0	51	0.3975	0.9
0.0921	52.0	52	0.4261	0.8333
0.0921	53.0	53	0.3944	0.8
0.0921	54.0	54	0.2987	0.9333
0.0921	55.0	55	0.4845	0.8667
0.0921	56.0	56	0.5880	0.7667
0.0921	57.0	57	0.6478	0.8333
0.0921	58.0	58	0.4498	0.8
0.0921	59.0	59	0.3165	0.8667
0.0488	60.0	60	0.5294	0.8333
0.0488	61.0	61	0.6030	0.8333
0.0488	62.0	62	0.4018	0.8333
0.0488	63.0	63	0.5076	0.8333
0.0488	64.0	64	0.5128	0.8667
0.0488	65.0	65	0.5164	0.8667
0.0488	66.0	66	0.4238	0.8333
0.0488	67.0	67	0.5057	0.8333
0.0488	68.0	68	0.6507	0.7667
0.0488	69.0	69	0.4623	0.8667
0.0336	70.0	70	0.4230	0.8333
0.0336	71.0	71	0.4669	0.8333
0.0336	72.0	72	0.4836	0.8333
0.0336	73.0	73	0.3458	0.9333
0.0336	74.0	74	0.4629	0.8667
0.0336	75.0	75	0.4426	0.7667
0.0336	76.0	76	0.4735	0.8
0.0336	77.0	77	0.5138	0.7667
0.0336	78.0	78	0.4728	0.8333
0.0336	79.0	79	0.3224	0.8667
0.0204	80.0	80	0.2733	0.8667
0.0204	81.0	81	0.4948	0.8333
0.0204	82.0	82	0.3923	0.9
0.0204	83.0	83	0.2380	0.9
0.0204	84.0	84	0.4343	0.8667
0.0204	85.0	85	0.4008	0.8
0.0204	86.0	86	0.3960	0.9
0.0204	87.0	87	0.4185	0.8667
0.0204	88.0	88	0.4394	0.8
0.0204	89.0	89	0.3055	0.9
0.0113	90.0	90	0.4782	0.7333
0.0113	91.0	91	0.4763	0.8667
0.0113	92.0	92	0.4404	0.9
0.0113	93.0	93	0.2787	0.9
0.0113	94.0	94	0.3599	0.9
0.0113	95.0	95	0.5665	0.8333
0.0113	96.0	96	0.3193	0.9333
0.0113	97.0	97	0.3259	0.8667
0.0113	98.0	98	0.3528	0.9333
0.0113	99.0	99	0.3905	0.8667
0.009	100.0	100	0.4000	0.8667

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0