hushem_40x_deit_tiny_sgd_0001_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: 1.1471
• Accuracy: 0.5333

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
• 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
1.4892	1.0	215	1.3892	0.2222
1.3775	2.0	430	1.3751	0.2889
1.3266	3.0	645	1.3631	0.3111
1.2619	4.0	860	1.3523	0.3111
1.235	5.0	1075	1.3429	0.3111
1.1826	6.0	1290	1.3347	0.3778
1.2015	7.0	1505	1.3284	0.3778
1.2072	8.0	1720	1.3225	0.4
1.1254	9.0	1935	1.3170	0.4
1.1293	10.0	2150	1.3118	0.3556
1.0925	11.0	2365	1.3069	0.3778
1.0731	12.0	2580	1.3024	0.3778
1.0421	13.0	2795	1.2976	0.3778
1.0531	14.0	3010	1.2928	0.3778
1.0284	15.0	3225	1.2877	0.3778
1.0283	16.0	3440	1.2824	0.4
1.0283	17.0	3655	1.2767	0.4222
1.0038	18.0	3870	1.2712	0.4444
0.9952	19.0	4085	1.2654	0.4444
0.9413	20.0	4300	1.2587	0.4444
0.9562	21.0	4515	1.2529	0.4667
1.0163	22.0	4730	1.2467	0.4667
0.9391	23.0	4945	1.2401	0.4667
0.955	24.0	5160	1.2340	0.4889
0.9454	25.0	5375	1.2281	0.4889
0.9013	26.0	5590	1.2229	0.4889
0.8818	27.0	5805	1.2169	0.5111
0.8594	28.0	6020	1.2115	0.5111
0.8984	29.0	6235	1.2064	0.5111
0.8277	30.0	6450	1.2009	0.5111
0.8636	31.0	6665	1.1955	0.5111
0.8466	32.0	6880	1.1910	0.5111
0.8955	33.0	7095	1.1866	0.5111
0.817	34.0	7310	1.1825	0.5111
0.8132	35.0	7525	1.1781	0.5111
0.7914	36.0	7740	1.1742	0.5111
0.835	37.0	7955	1.1705	0.5111
0.8383	38.0	8170	1.1668	0.5111
0.828	39.0	8385	1.1638	0.5111
0.7822	40.0	8600	1.1606	0.5111
0.8243	41.0	8815	1.1580	0.5333
0.9371	42.0	9030	1.1556	0.5333
0.8482	43.0	9245	1.1533	0.5333
0.8054	44.0	9460	1.1516	0.5333
0.8152	45.0	9675	1.1501	0.5333
0.8013	46.0	9890	1.1489	0.5333
0.7786	47.0	10105	1.1481	0.5333
0.7918	48.0	10320	1.1474	0.5333
0.8671	49.0	10535	1.1471	0.5333
0.8286	50.0	10750	1.1471	0.5333

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.1+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2