hushem_1x_deit_tiny_adamax_lr001_fold1

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: 1.5661
• Accuracy: 0.4444

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	2.2715	0.2667
No log	2.0	3	2.0269	0.4
No log	2.67	4	1.6111	0.2889
No log	4.0	6	1.4755	0.2444
No log	4.67	7	1.3818	0.4667
No log	6.0	9	1.3523	0.3111
1.6844	6.67	10	1.4010	0.2444
1.6844	8.0	12	1.2634	0.4444
1.6844	8.67	13	1.3983	0.4222
1.6844	10.0	15	1.7897	0.3778
1.6844	10.67	16	1.7305	0.3111
1.6844	12.0	18	1.3560	0.4667
1.6844	12.67	19	1.8545	0.4222
1.001	14.0	21	2.1000	0.3778
1.001	14.67	22	1.2257	0.4889
1.001	16.0	24	1.2741	0.4444
1.001	16.67	25	1.9098	0.3556
1.001	18.0	27	1.4981	0.3778
1.001	18.67	28	1.0949	0.4222
0.7366	20.0	30	1.1640	0.4222
0.7366	20.67	31	1.5156	0.3556
0.7366	22.0	33	1.8559	0.3556
0.7366	22.67	34	1.5735	0.4444
0.7366	24.0	36	1.3202	0.4222
0.7366	24.67	37	1.3837	0.4222
0.7366	26.0	39	1.6707	0.4
0.4908	26.67	40	1.8712	0.3778
0.4908	28.0	42	2.1885	0.3556
0.4908	28.67	43	2.0505	0.3556
0.4908	30.0	45	1.6855	0.4
0.4908	30.67	46	1.5304	0.4222
0.4908	32.0	48	1.5067	0.3778
0.4908	32.67	49	1.5442	0.4222
0.3287	33.33	50	1.5661	0.4444

Framework versions

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