Fer_vit_jaffe_GOOGLE_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: 0.4137
• Accuracy: 0.8333

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: 1
• 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	2.4981	0.1333
No log	2.0	2	2.4022	0.1333
No log	3.0	3	2.2167	0.1
No log	4.0	4	2.0743	0.1333
No log	5.0	5	1.9393	0.1
No log	6.0	6	2.0201	0.1667
No log	7.0	7	1.9793	0.1667
No log	8.0	8	1.9287	0.2
No log	9.0	9	1.8316	0.1667
2.1031	10.0	10	1.6923	0.4667
2.1031	11.0	11	1.7380	0.2667
2.1031	12.0	12	1.7164	0.3333
2.1031	13.0	13	1.6525	0.3333
2.1031	14.0	14	1.5759	0.3333
2.1031	15.0	15	1.5251	0.3333
2.1031	16.0	16	1.4557	0.4333
2.1031	17.0	17	1.3619	0.4333
2.1031	18.0	18	1.2880	0.4333
2.1031	19.0	19	1.2356	0.5667
1.2981	20.0	20	1.1369	0.6
1.2981	21.0	21	1.1489	0.5
1.2981	22.0	22	1.0756	0.7
1.2981	23.0	23	1.0136	0.5333
1.2981	24.0	24	1.0509	0.5333
1.2981	25.0	25	0.9975	0.6
1.2981	26.0	26	0.9895	0.6
1.2981	27.0	27	0.9735	0.6
1.2981	28.0	28	0.9328	0.6
1.2981	29.0	29	0.9559	0.6667
0.5735	30.0	30	0.8359	0.7667
0.5735	31.0	31	0.8023	0.7667
0.5735	32.0	32	0.8285	0.6333
0.5735	33.0	33	0.7287	0.7
0.5735	34.0	34	0.7043	0.7667
0.5735	35.0	35	0.8992	0.7333
0.5735	36.0	36	0.8664	0.7667
0.5735	37.0	37	0.8023	0.7333
0.5735	38.0	38	0.6910	0.7667
0.5735	39.0	39	0.8197	0.6667
0.2477	40.0	40	0.5915	0.7667
0.2477	41.0	41	0.9184	0.6333
0.2477	42.0	42	0.6734	0.7
0.2477	43.0	43	0.9225	0.7
0.2477	44.0	44	0.5961	0.8
0.2477	45.0	45	0.7012	0.7
0.2477	46.0	46	0.9223	0.6
0.2477	47.0	47	0.5819	0.7
0.2477	48.0	48	0.7171	0.7333
0.2477	49.0	49	0.6416	0.7667
0.1117	50.0	50	0.8718	0.7
0.1117	51.0	51	0.4941	0.8
0.1117	52.0	52	0.7385	0.8
0.1117	53.0	53	0.6660	0.8333
0.1117	54.0	54	0.6988	0.8667
0.1117	55.0	55	0.7074	0.7667
0.1117	56.0	56	0.5847	0.8
0.1117	57.0	57	0.6636	0.8
0.1117	58.0	58	0.5520	0.8333
0.1117	59.0	59	0.6299	0.7667
0.0591	60.0	60	0.6717	0.7667
0.0591	61.0	61	0.4874	0.8333
0.0591	62.0	62	0.4603	0.8
0.0591	63.0	63	0.5516	0.7333
0.0591	64.0	64	0.4729	0.8
0.0591	65.0	65	0.5710	0.7667
0.0591	66.0	66	0.8985	0.7
0.0591	67.0	67	0.8074	0.7667
0.0591	68.0	68	0.5652	0.8
0.0591	69.0	69	0.5538	0.8333
0.0296	70.0	70	0.5727	0.7333
0.0296	71.0	71	0.6359	0.8
0.0296	72.0	72	0.6932	0.7333
0.0296	73.0	73	0.9025	0.6667
0.0296	74.0	74	0.6639	0.7333
0.0296	75.0	75	0.8385	0.7333
0.0296	76.0	76	0.5827	0.8
0.0296	77.0	77	0.5443	0.8667
0.0296	78.0	78	0.6330	0.8333
0.0296	79.0	79	0.6706	0.7333
0.0175	80.0	80	0.7803	0.8
0.0175	81.0	81	0.5401	0.8
0.0175	82.0	82	0.6806	0.8333
0.0175	83.0	83	0.3827	0.8
0.0175	84.0	84	0.7853	0.8
0.0175	85.0	85	0.4391	0.8333
0.0175	86.0	86	0.6061	0.8
0.0175	87.0	87	0.4797	0.8
0.0175	88.0	88	0.4386	0.8333
0.0175	89.0	89	0.6556	0.8
0.0121	90.0	90	0.7927	0.8
0.0121	91.0	91	0.4925	0.8333
0.0121	92.0	92	0.6280	0.7667
0.0121	93.0	93	0.3561	0.9
0.0121	94.0	94	0.6058	0.8333
0.0121	95.0	95	0.5086	0.8
0.0121	96.0	96	0.3854	0.8667
0.0121	97.0	97	0.8370	0.7667
0.0121	98.0	98	0.6506	0.7333
0.0121	99.0	99	0.6100	0.7667
0.0088	100.0	100	0.4137	0.8333

Framework versions

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