emotion_classification_adjusted

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 0.8104
• Accuracy: 0.8875

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: 2e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 60
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Accuracy	Validation Loss
2.0787	1.0	20	0.1625	2.0753
2.073	2.0	40	0.1187	2.0737
2.0599	3.0	60	0.1938	2.0585
2.0363	4.0	80	0.1938	2.0368
2.0051	5.0	100	0.2625	1.9921
1.9348	6.0	120	0.3375	1.9185
1.8466	7.0	140	0.375	1.8056
1.755	8.0	160	0.4313	1.7292
1.676	9.0	180	0.45	1.6674
1.6244	10.0	200	0.475	1.6237
1.5661	11.0	220	0.5062	1.5973
1.5252	12.0	240	0.5	1.5262
1.4729	13.0	260	0.55	1.5050
1.4203	14.0	280	0.55	1.4784
1.364	15.0	300	0.525	1.5131
1.3262	16.0	320	0.5125	1.4776
1.3102	17.0	340	0.5563	1.4200
1.2595	18.0	360	0.5563	1.4329
1.2188	19.0	380	0.5375	1.4213
1.1991	20.0	400	0.525	1.4077
1.1526	21.0	420	0.6062	1.3625
1.1225	22.0	440	0.5437	1.3745
1.1283	23.0	460	0.5375	1.3677
1.0856	24.0	480	0.5625	1.3283
1.0559	25.0	500	0.5687	1.3440
1.0102	26.0	520	0.5437	1.3357
0.9915	27.0	540	0.5813	1.3377
0.9807	28.0	560	0.55	1.3824
0.9382	29.0	580	0.4938	1.4468
0.9857	30.0	600	0.8125	0.9923
0.9956	31.0	620	0.7625	1.0361
0.9875	32.0	640	0.775	1.0310
0.9582	33.0	660	0.7625	1.0572
0.9649	34.0	680	0.8063	0.9725
0.9099	35.0	700	0.7562	1.0355
0.9339	36.0	720	0.7937	1.0129
0.9045	37.0	740	0.7562	1.0315
0.8903	38.0	760	0.8187	0.9923
0.8799	39.0	780	0.7625	1.0386
0.8664	40.0	800	0.7438	1.0626
0.8351	41.0	820	0.7688	0.9885
0.8514	42.0	840	0.7875	0.9975
0.857	43.0	860	0.75	1.0169
0.8331	44.0	880	0.7937	0.9763
0.8093	45.0	900	0.7937	0.9645
0.8303	46.0	920	0.8	0.9880
0.8077	47.0	940	0.8063	1.0094
0.8082	48.0	960	0.7937	0.9757
0.8088	49.0	980	0.7438	1.0451
0.7985	50.0	1000	0.7875	0.9850
0.8013	51.0	1020	0.7688	1.0362
0.7882	52.0	1040	0.775	1.0007
0.8051	53.0	1060	0.7438	1.0314
0.812	54.0	1080	0.8	0.9782
0.7895	55.0	1100	0.725	1.0396
0.8012	56.0	1120	0.7688	0.9894
0.7973	57.0	1140	0.7875	0.9981
0.7946	58.0	1160	0.8063	0.9754
0.8437	59.0	1180	0.85	0.8544
0.8489	60.0	1200	0.7991	0.9062

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1
• Datasets 3.2.0
• Tokenizers 0.21.0