smids_3x_deit_tiny_adamax_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: 1.3554
• Accuracy: 0.875

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
0.3155	1.0	225	0.3767	0.855
0.1407	2.0	450	0.3968	0.87
0.0909	3.0	675	0.5151	0.8617
0.1097	4.0	900	0.4932	0.8633
0.0371	5.0	1125	0.6275	0.8683
0.0101	6.0	1350	0.7429	0.8633
0.0391	7.0	1575	0.8418	0.865
0.0077	8.0	1800	0.8293	0.8767
0.0687	9.0	2025	0.9529	0.8667
0.0191	10.0	2250	1.0880	0.8633
0.0325	11.0	2475	1.1276	0.8583
0.0003	12.0	2700	1.0880	0.8583
0.0043	13.0	2925	1.1462	0.8717
0.0003	14.0	3150	1.2942	0.8467
0.04	15.0	3375	1.1259	0.86
0.0007	16.0	3600	1.1878	0.86
0.0001	17.0	3825	1.2005	0.8683
0.0003	18.0	4050	1.1961	0.8667
0.0	19.0	4275	1.1734	0.8717
0.0002	20.0	4500	1.2441	0.865
0.0	21.0	4725	1.2631	0.87
0.0	22.0	4950	1.1604	0.8733
0.0	23.0	5175	1.1940	0.8683
0.0	24.0	5400	1.2754	0.8683
0.0	25.0	5625	1.2790	0.8617
0.0	26.0	5850	1.2664	0.8733
0.0	27.0	6075	1.2877	0.875
0.0	28.0	6300	1.2788	0.88
0.0	29.0	6525	1.2589	0.8783
0.0	30.0	6750	1.2671	0.8783
0.0	31.0	6975	1.2750	0.8783
0.0	32.0	7200	1.2783	0.875
0.0	33.0	7425	1.2973	0.8733
0.0	34.0	7650	1.3104	0.8717
0.0	35.0	7875	1.3108	0.8717
0.0	36.0	8100	1.3246	0.8717
0.0	37.0	8325	1.3217	0.87
0.0	38.0	8550	1.3349	0.8733
0.0	39.0	8775	1.3355	0.8733
0.0	40.0	9000	1.3337	0.8733
0.0	41.0	9225	1.3326	0.875
0.0036	42.0	9450	1.3380	0.8733
0.0	43.0	9675	1.3430	0.875
0.0032	44.0	9900	1.3445	0.8733
0.0	45.0	10125	1.3492	0.875
0.0	46.0	10350	1.3514	0.875
0.0	47.0	10575	1.3535	0.875
0.0	48.0	10800	1.3548	0.875
0.0	49.0	11025	1.3543	0.875
0.0	50.0	11250	1.3554	0.875

Framework versions

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