# Marine Life Identification

SegFormer model optimized for coral reef semantic segmentation tasks, fine-tuned on the Coralscapes dataset at 1024x1024 resolution

zoalearn2 is a coral species classification model fine-tuned based on Google's ViT model, capable of identifying 32 of the most popular coral species.

This is an image classification model for identifying and classifying marine mammals, with an accuracy rate of 84.72%.