distill-lab/train_dit02_3cat_cls_5.6k_naflex

train_dit02_initials_3cat_cls_5.6k_naflex-r2

This model is a fine-tuned version of google/siglip2-so400m-patch16-naflex on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.6709
• Accuracy: 0.6333
• Precision: 0.6313
• Recall: 0.6333
• F1: 0.6300
• Roc Auc: 0.7773

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: 5e-06
• train_batch_size: 96
• eval_batch_size: 96
• seed: 1337
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 10.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Precision	Recall	F1	Roc Auc
0.3921	1.8868	100	0.4369	0.6243	0.5741	0.6243	0.5256	0.7611
0.2374	3.7736	200	0.4151	0.6558	0.6163	0.6558	0.6183	0.7908
0.1115	5.6604	300	0.5398	0.6513	0.6193	0.6513	0.6309	0.7897
0.0411	7.5472	400	0.6090	0.6277	0.6382	0.6277	0.6289	0.7803
0.026	9.4340	500	0.6684	0.6333	0.6330	0.6333	0.6307	0.7774

Framework versions

• Transformers 4.56.2
• Pytorch 2.7.1+cu126
• Datasets 4.0.0
• Tokenizers 0.22.1