vit-ena24-MD

This model is a fine-tuned version of google/vit-base-patch16-224-in21k on the ena24_MD dataset. It achieves the following results on the evaluation set:

• Loss: 1.6827
• Accuracy: 0.6826

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• num_epochs: 2
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.4013	0.1259	100	1.8994	0.6631
0.4178	0.2519	200	2.0700	0.5938
0.336	0.3778	300	1.6827	0.6826
0.2547	0.5038	400	1.8338	0.6582
0.1166	0.6297	500	1.9549	0.6699
0.0555	0.7557	600	1.7759	0.7021
0.3521	0.8816	700	2.0155	0.6592
0.0144	1.0076	800	1.9693	0.6738
0.0847	1.1335	900	1.7930	0.7227
0.0041	1.2594	1000	1.7101	0.7334
0.0232	1.3854	1100	1.7127	0.7324
0.1154	1.5113	1200	1.8169	0.7236
0.0355	1.6373	1300	1.7981	0.7334
0.0042	1.7632	1400	1.7519	0.7451
0.0435	1.8892	1500	1.8185	0.7344

Framework versions

• Transformers 4.51.3
• Pytorch 2.6.0+cu124
• Datasets 3.6.0
• Tokenizers 0.21.1