hushem_1x_deit_tiny_adamax_lr00001_fold5

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: 1.2718
• Accuracy: 0.4634

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	1.6411	0.1220
No log	2.0	3	1.5318	0.1951
No log	2.67	4	1.5063	0.1707
No log	4.0	6	1.4764	0.3171
No log	4.67	7	1.4625	0.3171
No log	6.0	9	1.4352	0.3902
1.379	6.67	10	1.4210	0.4390
1.379	8.0	12	1.3993	0.4390
1.379	8.67	13	1.3906	0.4390
1.379	10.0	15	1.3747	0.4146
1.379	10.67	16	1.3676	0.4146
1.379	12.0	18	1.3554	0.4146
1.379	12.67	19	1.3500	0.4146
1.107	14.0	21	1.3389	0.4146
1.107	14.67	22	1.3348	0.4146
1.107	16.0	24	1.3265	0.4390
1.107	16.67	25	1.3236	0.4634
1.107	18.0	27	1.3162	0.4634
1.107	18.67	28	1.3129	0.4390
0.9495	20.0	30	1.3051	0.4390
0.9495	20.67	31	1.3019	0.4390
0.9495	22.0	33	1.2961	0.4390
0.9495	22.67	34	1.2934	0.4634
0.9495	24.0	36	1.2879	0.4390
0.9495	24.67	37	1.2851	0.4390
0.9495	26.0	39	1.2815	0.4390
0.8401	26.67	40	1.2802	0.4390
0.8401	28.0	42	1.2775	0.4390
0.8401	28.67	43	1.2761	0.4390
0.8401	30.0	45	1.2740	0.4390
0.8401	30.67	46	1.2734	0.4390
0.8401	32.0	48	1.2723	0.4634
0.8401	32.67	49	1.2719	0.4634
0.7816	33.33	50	1.2718	0.4634

Framework versions

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