hushem_1x_deit_tiny_adamax_lr0001_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: 0.5599
• 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: 0.0001
• 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	1.6749	0.3095
No log	2.0	3	1.3545	0.3333
No log	2.67	4	1.3451	0.2857
No log	4.0	6	1.2535	0.5238
No log	4.67	7	1.2290	0.4286
No log	6.0	9	1.1555	0.5
1.2457	6.67	10	1.0938	0.5
1.2457	8.0	12	0.9608	0.4762
1.2457	8.67	13	0.8825	0.5952
1.2457	10.0	15	0.7678	0.7143
1.2457	10.67	16	0.7184	0.7857
1.2457	12.0	18	0.6658	0.7619
1.2457	12.67	19	0.6361	0.7619
0.4167	14.0	21	0.6247	0.8095
0.4167	14.67	22	0.6111	0.7857
0.4167	16.0	24	0.5896	0.7857
0.4167	16.67	25	0.5886	0.7381
0.4167	18.0	27	0.6107	0.7619
0.4167	18.67	28	0.6198	0.7619
0.0627	20.0	30	0.6194	0.7619
0.0627	20.67	31	0.6092	0.7619
0.0627	22.0	33	0.5917	0.7857
0.0627	22.67	34	0.5871	0.7857
0.0627	24.0	36	0.5872	0.8095
0.0627	24.67	37	0.5896	0.8095
0.0627	26.0	39	0.5921	0.8095
0.0081	26.67	40	0.5908	0.8095
0.0081	28.0	42	0.5818	0.8095
0.0081	28.67	43	0.5772	0.8095
0.0081	30.0	45	0.5685	0.8095
0.0081	30.67	46	0.5654	0.8095
0.0081	32.0	48	0.5614	0.8333
0.0081	32.67	49	0.5603	0.8333
0.0038	33.33	50	0.5599	0.8333

Framework versions

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