smids_3x_deit_small_adamax_001_fold4

This model is a fine-tuned version of facebook/deit-small-patch16-224 on the imagefolder dataset. It achieves the following results on the evaluation set:

• Loss: 1.4399
• Accuracy: 0.86

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
• 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
0.3979	1.0	225	0.5683	0.7933
0.2606	2.0	450	0.4043	0.84
0.2162	3.0	675	0.5101	0.82
0.2443	4.0	900	0.4744	0.84
0.1743	5.0	1125	0.4964	0.8567
0.1364	6.0	1350	0.5063	0.8433
0.1949	7.0	1575	0.5298	0.855
0.116	8.0	1800	0.6753	0.8617
0.1838	9.0	2025	0.6150	0.84
0.0967	10.0	2250	0.7712	0.8433
0.0509	11.0	2475	0.7887	0.8417
0.0235	12.0	2700	0.6823	0.8633
0.0623	13.0	2925	0.6980	0.8533
0.0443	14.0	3150	0.6897	0.87
0.0341	15.0	3375	0.8718	0.8417
0.037	16.0	3600	0.8120	0.865
0.0314	17.0	3825	0.8372	0.8467
0.0477	18.0	4050	0.7518	0.85
0.0002	19.0	4275	1.0178	0.8517
0.0091	20.0	4500	1.1728	0.8333
0.0291	21.0	4725	0.9139	0.85
0.0076	22.0	4950	1.0132	0.8533
0.0005	23.0	5175	1.0336	0.8567
0.0002	24.0	5400	1.1694	0.85
0.0032	25.0	5625	1.1362	0.86
0.0002	26.0	5850	1.2122	0.85
0.011	27.0	6075	1.2712	0.8567
0.0237	28.0	6300	1.3743	0.85
0.0	29.0	6525	1.2063	0.86
0.0	30.0	6750	1.3085	0.86
0.0	31.0	6975	1.3297	0.8567
0.0	32.0	7200	1.2473	0.855
0.0	33.0	7425	1.1982	0.8617
0.0	34.0	7650	1.2288	0.8617
0.0	35.0	7875	1.2397	0.8617
0.0	36.0	8100	1.2697	0.8617
0.0	37.0	8325	1.2895	0.8583
0.0	38.0	8550	1.3064	0.8583
0.0	39.0	8775	1.3040	0.8583
0.0	40.0	9000	1.3224	0.8617
0.0	41.0	9225	1.3481	0.8567
0.0032	42.0	9450	1.3550	0.8583
0.0	43.0	9675	1.3631	0.86
0.0029	44.0	9900	1.3898	0.8567
0.0	45.0	10125	1.3948	0.86
0.0	46.0	10350	1.4079	0.86
0.0	47.0	10575	1.4219	0.86
0.0	48.0	10800	1.4319	0.8583
0.0	49.0	11025	1.4436	0.8583
0.0	50.0	11250	1.4399	0.86

Framework versions

• Transformers 4.32.1
• Pytorch 2.1.0+cu121
• Datasets 2.12.0
• Tokenizers 0.13.2