hushem_1x_deit_tiny_adamax_lr00001_fold2

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: 1.4306
• Accuracy: 0.2889

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 128
• 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	0.67	1	1.6419	0.1111
No log	2.0	3	1.5215	0.1556
No log	2.67	4	1.5000	0.1778
No log	4.0	6	1.4887	0.2444
No log	4.67	7	1.4849	0.2222
No log	6.0	9	1.4754	0.2444
1.3642	6.67	10	1.4684	0.2667
1.3642	8.0	12	1.4571	0.2667
1.3642	8.67	13	1.4523	0.2667
1.3642	10.0	15	1.4422	0.2667
1.3642	10.67	16	1.4392	0.2444
1.3642	12.0	18	1.4341	0.2444
1.3642	12.67	19	1.4327	0.2444
1.1012	14.0	21	1.4319	0.2667
1.1012	14.67	22	1.4329	0.2667
1.1012	16.0	24	1.4330	0.2889
1.1012	16.67	25	1.4333	0.2889
1.1012	18.0	27	1.4342	0.2889
1.1012	18.67	28	1.4339	0.2889
0.9232	20.0	30	1.4351	0.2889
0.9232	20.67	31	1.4354	0.2889
0.9232	22.0	33	1.4352	0.2889
0.9232	22.67	34	1.4353	0.2889
0.9232	24.0	36	1.4349	0.2889
0.9232	24.67	37	1.4347	0.2889
0.9232	26.0	39	1.4341	0.2889
0.8235	26.67	40	1.4334	0.2889
0.8235	28.0	42	1.4325	0.2889
0.8235	28.67	43	1.4325	0.2889
0.8235	30.0	45	1.4317	0.2889
0.8235	30.67	46	1.4312	0.2889
0.8235	32.0	48	1.4307	0.2889
0.8235	32.67	49	1.4307	0.2889
0.7752	33.33	50	1.4306	0.2889

Framework versions

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