hushem_40x_deit_small_sgd_0001_fold3

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: 1.0466
• Accuracy: 0.5814

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
1.6887	1.0	217	1.4610	0.2558
1.5962	2.0	434	1.3845	0.3488
1.5124	3.0	651	1.3560	0.3721
1.442	4.0	868	1.3419	0.3721
1.41	5.0	1085	1.3313	0.3488
1.3709	6.0	1302	1.3218	0.3721
1.3157	7.0	1519	1.3125	0.3721
1.3328	8.0	1736	1.3039	0.3488
1.3107	9.0	1953	1.2950	0.3488
1.2568	10.0	2170	1.2861	0.3488
1.2226	11.0	2387	1.2769	0.3256
1.198	12.0	2604	1.2671	0.3256
1.232	13.0	2821	1.2570	0.3488
1.1803	14.0	3038	1.2472	0.3488
1.214	15.0	3255	1.2376	0.3488
1.208	16.0	3472	1.2274	0.3953
1.1406	17.0	3689	1.2176	0.3953
1.1243	18.0	3906	1.2072	0.3953
1.1316	19.0	4123	1.1970	0.4884
1.1119	20.0	4340	1.1873	0.4884
1.117	21.0	4557	1.1775	0.5116
1.0609	22.0	4774	1.1681	0.5116
1.0751	23.0	4991	1.1588	0.5581
1.058	24.0	5208	1.1499	0.5581
1.0301	25.0	5425	1.1417	0.5581
1.089	26.0	5642	1.1338	0.5581
0.9909	27.0	5859	1.1255	0.5814
0.9932	28.0	6076	1.1180	0.5814
1.026	29.0	6293	1.1110	0.5814
1.0236	30.0	6510	1.1044	0.5814
1.0169	31.0	6727	1.0980	0.5814
1.0049	32.0	6944	1.0921	0.5814
1.0261	33.0	7161	1.0868	0.5814
0.994	34.0	7378	1.0819	0.5814
0.9887	35.0	7595	1.0769	0.5581
1.0137	36.0	7812	1.0725	0.5581
0.9359	37.0	8029	1.0687	0.5581
0.9531	38.0	8246	1.0651	0.5581
0.9682	39.0	8463	1.0620	0.5581
0.9947	40.0	8680	1.0590	0.5581
0.9063	41.0	8897	1.0565	0.5581
1.0195	42.0	9114	1.0543	0.5581
0.966	43.0	9331	1.0523	0.5581
0.9409	44.0	9548	1.0506	0.5581
0.9327	45.0	9765	1.0492	0.5581
0.9575	46.0	9982	1.0481	0.5814
0.9627	47.0	10199	1.0474	0.5814
0.9553	48.0	10416	1.0469	0.5814
0.9631	49.0	10633	1.0467	0.5814
0.944	50.0	10850	1.0466	0.5814

Framework versions

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