id_ravdess_mel_spec_Vit_vit-tiny-patch16-224_1

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: 0.8036
• Accuracy: 0.7917

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: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	1.0	8	3.5290	0.0231
3.7813	2.0	16	3.1577	0.0880
3.1834	3.0	24	2.9675	0.1435
2.8849	4.0	32	2.6636	0.2037
2.3044	5.0	40	2.2016	0.3565
2.3044	6.0	48	1.8205	0.4444
1.5789	7.0	56	1.5203	0.5417
1.0355	8.0	64	1.3309	0.6481
0.5941	9.0	72	1.2203	0.6528
0.3301	10.0	80	1.1015	0.6806
0.3301	11.0	88	1.0155	0.7176
0.1632	12.0	96	0.9678	0.7454
0.0791	13.0	104	0.9265	0.7454
0.0347	14.0	112	0.8884	0.75
0.0165	15.0	120	0.8745	0.7454
0.0165	16.0	128	0.8517	0.7824
0.007	17.0	136	0.8396	0.7593
0.0033	18.0	144	0.8265	0.7917
0.0019	19.0	152	0.8109	0.7685
0.0013	20.0	160	0.8015	0.7917
0.0013	21.0	168	0.8209	0.7917
0.001	22.0	176	0.7898	0.7917
0.0008	23.0	184	0.8016	0.7824
0.0007	24.0	192	0.8024	0.7824
0.0006	25.0	200	0.8014	0.7917
0.0006	26.0	208	0.7987	0.7824
0.0006	27.0	216	0.8025	0.7870
0.0006	28.0	224	0.8009	0.7824
0.0005	29.0	232	0.8001	0.7824
0.0005	30.0	240	0.8014	0.7870
0.0005	31.0	248	0.8025	0.7824
0.0005	32.0	256	0.8013	0.7824
0.0004	33.0	264	0.8026	0.7824
0.0004	34.0	272	0.8004	0.7824
0.0004	35.0	280	0.8027	0.7870
0.0004	36.0	288	0.8011	0.7824
0.0004	37.0	296	0.8036	0.7917
0.0004	38.0	304	0.8027	0.7824
0.0004	39.0	312	0.8028	0.7870
0.0004	40.0	320	0.8031	0.7824
0.0004	41.0	328	0.8037	0.7824
0.0004	42.0	336	0.8028	0.7870
0.0003	43.0	344	0.8027	0.7870
0.0003	44.0	352	0.8028	0.7824
0.0003	45.0	360	0.8034	0.7870
0.0003	46.0	368	0.8032	0.7870
0.0003	47.0	376	0.8031	0.7917
0.0003	48.0	384	0.8034	0.7870
0.0003	49.0	392	0.8037	0.7917
0.0003	50.0	400	0.8036	0.7917

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.3.1
• Tokenizers 0.21.0