eyegazer-vit-lora / README.md
monsoon-nlp's picture
Update README.md
2262d47
metadata
library_name: peft

LoRA Image Classification LoRA adapter

Model: google/vit-base-patch16-224-in21k

Dataset: https://www.kaggle.com/c/aptos2019-blindness-detection - fundus images of the back of the eye, and diabetic retinopathy score

Source docs: https://huggingface.co/docs/peft/task_guides/image_classification_lora

Training notebook: https://colab.research.google.com/drive/1TVsUyyou87E26Sz40CdBH3CzWoVckgtq?usp=sharing

On 10% held-out of training data: accuracy 76%

  • PEFT 0.5.0

Future goals

  • More documentation
  • Modify loss for regression on 0-4 score
  • Script and Gradio to use raw images