links_to_notebooks.txt

Notebook: effnet-b5-complete_preprocessing-transfer_b5_old_new.ipynb
https://www.kaggle.com/yogendrayatnalkar/m1-effnetb5-using-ram-fork1/data?scriptVersionId=31299754
Description: The above model is first trained on old-dataseet without any preprocessing. Later with transfer learning the same model is trained on current aptos dataset with various image processing and zooming. This model is a bit over-fitted.

Notebook: ensemble.ipynb
https://www.kaggle.com/yogendrayatnalkar/fork-of-aptos-submission
Description: The above file shows how the 4 models are combined to give an edge over the accuracy.

Notebook: densenet-using-zoom.ipynb
https://www.kaggle.com/yogendrayatnalkar/densenet-using-zoom?scriptVersionId=31117820
Description: Used densenet as the base model. Tried the zoom effect on images.

Notebook: effnet-b5-new.ipynb
https://www.kaggle.com/yogendrayatnalkar/m1-effnetb5-using-ram-fork1?scriptVersionId=31116354
Description: Used standard efficient-net B5 on the current data itself.

Notebook: effnet-b3-new.ipynb
https://www.kaggle.com/hanozd1234/m1-effnetb3-using-ram
Description: Used standard efficient-net B3 on the current data itself.

Notebook: effnet-b5-old-new.ipynb
https://www.kaggle.com/hanozd1234/m1-effnetb5-using-ram-fork1-3a0430
Description: Model got initially trained in 10500 images of old data-set and then it was trained on new dataset with efficient-net B5.

Notebook: submission_checker.ipynb
https://www.kaggle.com/yogendrayatnalkar/aptos-submission/data?scriptVersionId=31338024
Description: To check the score of each individual model.