hushem_1x_deit_tiny_adamax_lr001_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: 1.4814
• Accuracy: 0.5333

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
• 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	4.5182	0.2444
No log	2.0	3	1.5416	0.2444
No log	2.67	4	1.5662	0.2667
No log	4.0	6	1.4453	0.2444
No log	4.67	7	1.4082	0.2444
No log	6.0	9	1.3188	0.4222
1.9051	6.67	10	1.3266	0.3556
1.9051	8.0	12	1.2375	0.4667
1.9051	8.67	13	1.3632	0.3778
1.9051	10.0	15	1.2064	0.4
1.9051	10.67	16	1.5392	0.2889
1.9051	12.0	18	1.1260	0.4889
1.9051	12.67	19	1.0999	0.4667
1.1808	14.0	21	1.2445	0.4222
1.1808	14.67	22	1.2069	0.4444
1.1808	16.0	24	1.0381	0.4889
1.1808	16.67	25	1.0992	0.5111
1.1808	18.0	27	1.1085	0.5333
1.1808	18.67	28	1.0609	0.5111
0.899	20.0	30	1.1754	0.5333
0.899	20.67	31	1.1214	0.5333
0.899	22.0	33	1.2625	0.4889
0.899	22.67	34	1.2586	0.5111
0.899	24.0	36	1.3423	0.4667
0.899	24.67	37	1.4290	0.4667
0.899	26.0	39	1.3722	0.5333
0.4924	26.67	40	1.4024	0.5111
0.4924	28.0	42	1.3396	0.5111
0.4924	28.67	43	1.4100	0.4444
0.4924	30.0	45	1.5561	0.4889
0.4924	30.67	46	1.5223	0.5556
0.4924	32.0	48	1.4581	0.5778
0.4924	32.67	49	1.4627	0.5556
0.1685	33.33	50	1.4814	0.5333

Framework versions

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