spycoder/vit-base-patch16-224-in21k-enhanced-ham10000

vit-base-beans

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the ahishamm/HAM_db_enhanced_balanced_reduced_50_20_20_50 dataset. It achieves the following results on the evaluation set:

• Loss: 0.5305
• Accuracy: 0.8451

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: 0.0002
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
1.0791	0.2304	100	1.0348	0.6335
0.9415	0.4608	200	0.9576	0.6449
0.7839	0.6912	300	0.8963	0.6662
0.7181	0.9217	400	0.8479	0.6963
0.3995	1.1521	500	0.7821	0.7170
0.5025	1.3825	600	0.6300	0.7837
0.4985	1.6129	700	0.7059	0.7490
0.4388	1.8433	800	0.5893	0.7857
0.2389	2.0737	900	0.5929	0.8077
0.2767	2.3041	1000	0.5795	0.8091
0.2387	2.5346	1100	0.6100	0.8091
0.1691	2.7650	1200	0.6175	0.8071
0.1738	2.9954	1300	0.5877	0.8198
0.0397	3.2258	1400	0.5766	0.8358
0.03	3.4562	1500	0.5681	0.8371
0.092	3.6866	1600	0.5305	0.8451
0.0416	3.9171	1700	0.5443	0.8471

Framework versions

• Transformers 4.41.2
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1