hushem_1x_deit_tiny_adamax_lr00001_fold1

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: 1.3005
• Accuracy: 0.4222

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	1.4838	0.2222
No log	2.0	3	1.4436	0.2222
No log	2.67	4	1.4334	0.1778
No log	4.0	6	1.4190	0.2667
No log	4.67	7	1.4121	0.2889
No log	6.0	9	1.3991	0.3111
1.3869	6.67	10	1.3926	0.3333
1.3869	8.0	12	1.3807	0.3556
1.3869	8.67	13	1.3748	0.3556
1.3869	10.0	15	1.3643	0.3778
1.3869	10.67	16	1.3598	0.3778
1.3869	12.0	18	1.3511	0.4
1.3869	12.67	19	1.3478	0.3778
1.1228	14.0	21	1.3405	0.4
1.1228	14.67	22	1.3380	0.4
1.1228	16.0	24	1.3323	0.4222
1.1228	16.67	25	1.3292	0.4222
1.1228	18.0	27	1.3250	0.4222
1.1228	18.67	28	1.3231	0.4222
0.9505	20.0	30	1.3201	0.4222
0.9505	20.67	31	1.3189	0.4222
0.9505	22.0	33	1.3162	0.4222
0.9505	22.67	34	1.3147	0.4222
0.9505	24.0	36	1.3120	0.4222
0.9505	24.67	37	1.3113	0.4222
0.9505	26.0	39	1.3090	0.4222
0.8411	26.67	40	1.3078	0.4222
0.8411	28.0	42	1.3057	0.4222
0.8411	28.67	43	1.3047	0.4222
0.8411	30.0	45	1.3028	0.4222
0.8411	30.67	46	1.3020	0.4222
0.8411	32.0	48	1.3010	0.4222
0.8411	32.67	49	1.3007	0.4222
0.7881	33.33	50	1.3005	0.4222

Framework versions

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