Comment by achristmascarl

For the model fine-tuned on biomedical image data, does anyone with domain knowledge know how the model's answers compare to the "Gold" answers?