hushem_1x_deit_tiny_adamax_lr001_fold3

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: 1.6696
• Accuracy: 0.5814

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: 0.001
• 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	4.2125	0.2558
No log	2.0	3	1.4682	0.2558
No log	2.67	4	1.6910	0.2558
No log	4.0	6	1.4476	0.2558
No log	4.67	7	1.3895	0.2558
No log	6.0	9	1.3751	0.2558
1.9073	6.67	10	1.3741	0.3953
1.9073	8.0	12	1.3957	0.3488
1.9073	8.67	13	1.3369	0.4419
1.9073	10.0	15	1.2847	0.4186
1.9073	10.67	16	1.3400	0.3953
1.9073	12.0	18	1.2676	0.3953
1.9073	12.67	19	1.2806	0.3721
1.1656	14.0	21	1.3652	0.3023
1.1656	14.67	22	1.3370	0.4419
1.1656	16.0	24	1.5165	0.3721
1.1656	16.67	25	1.5828	0.3953
1.1656	18.0	27	1.3210	0.3953
1.1656	18.67	28	1.3473	0.4419
0.9249	20.0	30	1.4346	0.4651
0.9249	20.67	31	1.3840	0.3953
0.9249	22.0	33	1.3578	0.4884
0.9249	22.67	34	1.3339	0.4884
0.9249	24.0	36	1.3509	0.4884
0.9249	24.67	37	1.3931	0.4884
0.9249	26.0	39	1.5691	0.5116
0.5495	26.67	40	1.5953	0.5349
0.5495	28.0	42	1.6688	0.5814
0.5495	28.67	43	1.6795	0.5581
0.5495	30.0	45	1.6839	0.5814
0.5495	30.67	46	1.6666	0.5814
0.5495	32.0	48	1.6555	0.5814
0.5495	32.67	49	1.6646	0.5814
0.2333	33.33	50	1.6696	0.5814

Framework versions

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