identifier-deberta

This model is a fine-tuned version of microsoft/deberta-v3-base on data derived from the PLABA dataset. Training was performed over 3 epochs with learning rate 2e-5. The model achieves the following performance:

• Validation Loss: 0.112134
• Precision: 0.455793
• Recall: 0.379442
• F1: 0.414127
• Accuracy: 0.961042