medsiglip-448-surgwound-6label

This model is a fine-tuned version of google/medsiglip-448 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1756
• Auc Healing Status: 0.5705
• Auc Erythema: 0.3988
• Auc Edema: 0.4804
• Auc Infection Risk: 0.6988
• Auc Urgency: 0.6525
• Auc Exudate: 0.8323
• Roc Auc Macro: 0.6055

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-05
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 16
• total_train_batch_size: 64
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Auc Healing Status	Auc Erythema	Auc Edema	Auc Infection Risk	Auc Urgency	Auc Exudate	Roc Auc Macro
7.2930	1.0	8	1.2750	0.5084	0.3466	0.4039	0.5519	0.4766	0.5747	0.4770
9.2917	2.0	16	1.2186	0.5322	0.3636	0.4510	0.6	0.5350	0.6919	0.5290
7.1242	3.0	24	1.1841	0.5529	0.3791	0.4667	0.6654	0.6168	0.7798	0.5768
9.5327	4.0	32	1.1773	0.5653	0.3967	0.4765	0.6951	0.6470	0.8283	0.6015
8.5318	5.0	40	1.1756	0.5705	0.3988	0.4804	0.6988	0.6525	0.8323	0.6055

Framework versions

• Transformers 5.2.0
• Pytorch 2.9.0+cu126
• Datasets 4.0.0
• Tokenizers 0.22.2