smids_3x_deit_tiny_rms_0001_fold4

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: 1.5044
• 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: 0.0001
• 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
0.4145	1.0	225	0.4267	0.8217
0.2495	2.0	450	0.4206	0.8567
0.1493	3.0	675	0.5020	0.8333
0.1961	4.0	900	0.4855	0.8517
0.1126	5.0	1125	0.6130	0.8317
0.0879	6.0	1350	0.5053	0.8667
0.0843	7.0	1575	0.6579	0.88
0.17	8.0	1800	0.6072	0.8617
0.0806	9.0	2025	0.8260	0.8517
0.1058	10.0	2250	0.9813	0.8283
0.0764	11.0	2475	0.8684	0.8567
0.005	12.0	2700	0.9093	0.8617
0.0735	13.0	2925	0.8151	0.8583
0.0859	14.0	3150	0.8528	0.87
0.0467	15.0	3375	1.0959	0.8533
0.0058	16.0	3600	1.0784	0.8517
0.0495	17.0	3825	1.1159	0.8467
0.0916	18.0	4050	1.0811	0.87
0.0002	19.0	4275	1.0640	0.855
0.0104	20.0	4500	1.2921	0.845
0.0459	21.0	4725	0.9148	0.87
0.0016	22.0	4950	1.1604	0.86
0.0121	23.0	5175	1.0782	0.855
0.0453	24.0	5400	1.1706	0.8633
0.0028	25.0	5625	1.1993	0.8633
0.0029	26.0	5850	1.1563	0.855
0.0156	27.0	6075	1.2423	0.8533
0.0005	28.0	6300	1.1448	0.8683
0.0001	29.0	6525	1.2011	0.8567
0.0043	30.0	6750	1.2731	0.85
0.0	31.0	6975	1.2673	0.8567
0.0328	32.0	7200	1.2677	0.8517
0.0122	33.0	7425	1.2900	0.8567
0.0	34.0	7650	1.4452	0.855
0.0	35.0	7875	1.2011	0.8733
0.0	36.0	8100	1.2263	0.8583
0.0009	37.0	8325	1.2364	0.87
0.0	38.0	8550	1.2628	0.865
0.0003	39.0	8775	1.3917	0.86
0.0	40.0	9000	1.4138	0.8617
0.0	41.0	9225	1.4172	0.8617
0.0032	42.0	9450	1.3990	0.865
0.0	43.0	9675	1.4412	0.86
0.0033	44.0	9900	1.4073	0.8617
0.0	45.0	10125	1.5048	0.86
0.0	46.0	10350	1.4842	0.8617
0.0	47.0	10575	1.4970	0.86
0.0	48.0	10800	1.4970	0.86
0.0	49.0	11025	1.5025	0.86
0.0	50.0	11250	1.5044	0.86

Framework versions

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