smids_3x_deit_tiny_adamax_0001_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: 0.8810
• Accuracy: 0.8933

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
0.3187	1.0	225	0.2926	0.8683
0.1675	2.0	450	0.2434	0.9083
0.148	3.0	675	0.3249	0.8883
0.0849	4.0	900	0.2613	0.91
0.0858	5.0	1125	0.4160	0.895
0.0448	6.0	1350	0.5146	0.8833
0.0473	7.0	1575	0.6012	0.8833
0.0246	8.0	1800	0.5599	0.89
0.0437	9.0	2025	0.6206	0.8933
0.0184	10.0	2250	0.6714	0.9017
0.0009	11.0	2475	0.6631	0.9083
0.0431	12.0	2700	0.7764	0.8983
0.0005	13.0	2925	0.6164	0.9033
0.0002	14.0	3150	0.6308	0.9017
0.0025	15.0	3375	0.7289	0.8983
0.0001	16.0	3600	0.6634	0.905
0.0	17.0	3825	0.7636	0.9033
0.0233	18.0	4050	0.7494	0.905
0.027	19.0	4275	0.8179	0.8917
0.0	20.0	4500	0.8300	0.895
0.0	21.0	4725	0.8262	0.8967
0.0	22.0	4950	0.8472	0.8917
0.0	23.0	5175	0.7368	0.9067
0.0091	24.0	5400	0.7922	0.8983
0.0	25.0	5625	0.8707	0.9
0.0	26.0	5850	0.7645	0.9017
0.0	27.0	6075	0.8363	0.8983
0.0	28.0	6300	0.8125	0.905
0.0076	29.0	6525	0.7853	0.9067
0.0	30.0	6750	0.8267	0.8967
0.0	31.0	6975	0.8018	0.905
0.0	32.0	7200	0.8256	0.9
0.0059	33.0	7425	0.8776	0.8967
0.0	34.0	7650	0.8060	0.9
0.0032	35.0	7875	0.8635	0.9
0.0	36.0	8100	0.8389	0.895
0.005	37.0	8325	0.8643	0.895
0.0	38.0	8550	0.8458	0.8983
0.0	39.0	8775	0.8735	0.895
0.0	40.0	9000	0.8584	0.895
0.0	41.0	9225	0.8812	0.895
0.0	42.0	9450	0.8710	0.895
0.0032	43.0	9675	0.8774	0.895
0.0	44.0	9900	0.8731	0.895
0.0	45.0	10125	0.8725	0.8967
0.0	46.0	10350	0.8786	0.895
0.0	47.0	10575	0.8795	0.895
0.0	48.0	10800	0.8801	0.895
0.0	49.0	11025	0.8803	0.8933
0.0	50.0	11250	0.8810	0.8933

Framework versions

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