Fer_vit_jaffe_crop_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: 0.4370
• Accuracy: 0.9

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	2.4680	0.1333
No log	2.0	2	2.2791	0.2333
No log	3.0	3	2.2505	0.1667
No log	4.0	4	2.0650	0.1
No log	5.0	5	2.1205	0.0333
No log	6.0	6	2.0198	0.1
No log	7.0	7	2.0317	0.1333
No log	8.0	8	1.9863	0.2333
No log	9.0	9	1.9390	0.3
2.1093	10.0	10	1.8465	0.2667
2.1093	11.0	11	1.6948	0.4333
2.1093	12.0	12	1.6453	0.4333
2.1093	13.0	13	1.6213	0.3333
2.1093	14.0	14	1.6045	0.3667
2.1093	15.0	15	1.5593	0.4333
2.1093	16.0	16	1.5160	0.5
2.1093	17.0	17	1.5322	0.4667
2.1093	18.0	18	1.4750	0.5
2.1093	19.0	19	1.3553	0.5333
1.3827	20.0	20	1.2704	0.4667
1.3827	21.0	21	1.2823	0.4667
1.3827	22.0	22	1.3789	0.5333
1.3827	23.0	23	1.2368	0.5667
1.3827	24.0	24	1.0561	0.6
1.3827	25.0	25	1.2039	0.5333
1.3827	26.0	26	1.2061	0.5333
1.3827	27.0	27	0.9144	0.6333
1.3827	28.0	28	1.0374	0.6
1.3827	29.0	29	1.0670	0.6333
0.6174	30.0	30	1.0691	0.6667
0.6174	31.0	31	0.9445	0.6667
0.6174	32.0	32	0.8885	0.5667
0.6174	33.0	33	0.9647	0.6
0.6174	34.0	34	1.0187	0.5667
0.6174	35.0	35	0.9037	0.6333
0.6174	36.0	36	0.9069	0.6
0.6174	37.0	37	0.8999	0.6333
0.6174	38.0	38	0.6198	0.7667
0.6174	39.0	39	0.8034	0.6667
0.2248	40.0	40	0.9049	0.6667
0.2248	41.0	41	0.7231	0.6667
0.2248	42.0	42	0.6554	0.7
0.2248	43.0	43	0.6591	0.8
0.2248	44.0	44	0.7196	0.8
0.2248	45.0	45	0.7233	0.7
0.2248	46.0	46	0.6112	0.8
0.2248	47.0	47	0.4299	0.8667
0.2248	48.0	48	0.5479	0.8
0.2248	49.0	49	0.5996	0.8333
0.0773	50.0	50	0.6714	0.7333
0.0773	51.0	51	0.4989	0.8333
0.0773	52.0	52	0.4956	0.8667
0.0773	53.0	53	0.4367	0.8333
0.0773	54.0	54	0.4542	0.8333
0.0773	55.0	55	0.5991	0.8
0.0773	56.0	56	0.6906	0.7667
0.0773	57.0	57	0.6667	0.7333
0.0773	58.0	58	0.5142	0.8
0.0773	59.0	59	0.5593	0.8
0.035	60.0	60	0.7527	0.7
0.035	61.0	61	0.4706	0.8667
0.035	62.0	62	0.5345	0.8333
0.035	63.0	63	0.5804	0.7667
0.035	64.0	64	0.5549	0.7667
0.035	65.0	65	0.5665	0.8
0.035	66.0	66	0.3258	0.9333
0.035	67.0	67	0.4890	0.8333
0.035	68.0	68	0.4657	0.8333
0.035	69.0	69	0.6546	0.8
0.0192	70.0	70	0.4962	0.8667
0.0192	71.0	71	0.5801	0.8
0.0192	72.0	72	0.5365	0.8667
0.0192	73.0	73	0.3524	0.8667
0.0192	74.0	74	0.5291	0.8667
0.0192	75.0	75	0.4613	0.9333
0.0192	76.0	76	0.5031	0.8
0.0192	77.0	77	0.4986	0.8333
0.0192	78.0	78	0.6103	0.8
0.0192	79.0	79	0.5855	0.8333
0.0126	80.0	80	0.6136	0.7667
0.0126	81.0	81	0.5112	0.8667
0.0126	82.0	82	0.4770	0.8333
0.0126	83.0	83	0.4016	0.8667
0.0126	84.0	84	0.4946	0.8667
0.0126	85.0	85	0.5542	0.7667
0.0126	86.0	86	0.4037	0.8667
0.0126	87.0	87	0.4775	0.8
0.0126	88.0	88	0.5146	0.8333
0.0126	89.0	89	0.5603	0.7667
0.0072	90.0	90	0.5734	0.8
0.0072	91.0	91	0.5937	0.8
0.0072	92.0	92	0.5328	0.8
0.0072	93.0	93	0.4362	0.8667
0.0072	94.0	94	0.6317	0.7667
0.0072	95.0	95	0.4078	0.8667
0.0072	96.0	96	0.5680	0.8
0.0072	97.0	97	0.6209	0.8
0.0072	98.0	98	0.5360	0.8
0.0072	99.0	99	0.4784	0.8667
0.0093	100.0	100	0.4370	0.9

Framework versions

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