hushem_1x_deit_tiny_adamax_lr00001_fold4

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

• Loss: 1.1208
• Accuracy: 0.5952

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• 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: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
No log	0.67	1	1.5521	0.1429
No log	2.0	3	1.4205	0.2857
No log	2.67	4	1.3862	0.3571
No log	4.0	6	1.3478	0.5238
No log	4.67	7	1.3332	0.5238
No log	6.0	9	1.3093	0.5238
1.4089	6.67	10	1.2970	0.5476
1.4089	8.0	12	1.2777	0.5714
1.4089	8.67	13	1.2689	0.5714
1.4089	10.0	15	1.2544	0.5714
1.4089	10.67	16	1.2478	0.5714
1.4089	12.0	18	1.2338	0.5714
1.4089	12.67	19	1.2267	0.5714
1.1506	14.0	21	1.2124	0.5714
1.1506	14.67	22	1.2049	0.5714
1.1506	16.0	24	1.1908	0.5714
1.1506	16.67	25	1.1843	0.5952
1.1506	18.0	27	1.1717	0.5952
1.1506	18.67	28	1.1659	0.5952
0.986	20.0	30	1.1576	0.5952
0.986	20.67	31	1.1537	0.5952
0.986	22.0	33	1.1470	0.5952
0.986	22.67	34	1.1439	0.5952
0.986	24.0	36	1.1385	0.5714
0.986	24.67	37	1.1362	0.5952
0.986	26.0	39	1.1320	0.5952
0.8708	26.67	40	1.1301	0.5952
0.8708	28.0	42	1.1268	0.5952
0.8708	28.67	43	1.1256	0.5952
0.8708	30.0	45	1.1234	0.5952
0.8708	30.67	46	1.1226	0.5952
0.8708	32.0	48	1.1214	0.5952
0.8708	32.67	49	1.1210	0.5952
0.8182	33.33	50	1.1208	0.5952

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1