hushem_1x_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: 3.1234
• Accuracy: 0.5952

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
No log	1.0	6	1.5633	0.2381
1.7898	2.0	12	1.3816	0.2381
1.7898	3.0	18	1.3607	0.2619
1.4334	4.0	24	1.3501	0.2619
1.3732	5.0	30	1.3553	0.2381
1.3732	6.0	36	1.1841	0.4762
1.3036	7.0	42	1.0576	0.5952
1.3036	8.0	48	1.0689	0.5952
1.2142	9.0	54	1.2296	0.5
1.056	10.0	60	0.7879	0.6429
1.056	11.0	66	0.7199	0.7143
0.921	12.0	72	0.9775	0.6190
0.921	13.0	78	0.8809	0.5952
0.6456	14.0	84	1.0792	0.5476
0.6348	15.0	90	1.0335	0.6190
0.6348	16.0	96	1.7853	0.5714
0.4743	17.0	102	1.5872	0.5714
0.4743	18.0	108	2.0651	0.5
0.2408	19.0	114	2.8369	0.4762
0.2271	20.0	120	2.1149	0.6190
0.2271	21.0	126	1.5722	0.6190
0.3385	22.0	132	2.8555	0.5476
0.3385	23.0	138	2.2068	0.6667
0.0822	24.0	144	2.2969	0.6190
0.0932	25.0	150	1.8785	0.7143
0.0932	26.0	156	3.2275	0.5714
0.0807	27.0	162	2.8847	0.5952
0.0807	28.0	168	3.1184	0.5952
0.0424	29.0	174	2.4583	0.6190
0.0287	30.0	180	2.8305	0.5714
0.0287	31.0	186	3.5171	0.5476
0.0333	32.0	192	3.2119	0.5952
0.0333	33.0	198	2.9811	0.5952
0.0008	34.0	204	3.0451	0.5952
0.0004	35.0	210	3.0670	0.5952
0.0004	36.0	216	3.0857	0.5952
0.0003	37.0	222	3.1009	0.5952
0.0003	38.0	228	3.1113	0.5952
0.0003	39.0	234	3.1177	0.5952
0.0003	40.0	240	3.1213	0.5952
0.0003	41.0	246	3.1231	0.5952
0.0002	42.0	252	3.1234	0.5952
0.0002	43.0	258	3.1234	0.5952
0.0002	44.0	264	3.1234	0.5952
0.0002	45.0	270	3.1234	0.5952
0.0002	46.0	276	3.1234	0.5952
0.0002	47.0	282	3.1234	0.5952
0.0002	48.0	288	3.1234	0.5952
0.0002	49.0	294	3.1234	0.5952
0.0002	50.0	300	3.1234	0.5952

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1