Topic recognition

[LargePenguin034]

Improved Unet model that is used to map a mask to skin Lessions in the ISIC dataset

Contains

dataset.py - Function for data loading
modules.py - Improved Unet model
predict.py - Used to use model after training
train.py - Used to train model

Aaron Reynolds
Student ID: 47479049

[wangzhaomxy]

This is an initial inspection, no action is required at this point.

• Files are a bit chaos, but containing all necessary ones. Trained model is included in the repository, which is not allowed as per the spec.

• Algorithm solves the problem appropriately. Average Dice Coefficient in testing dataset: 0.88, but no evidence.

• Correctly using PyTorch to construct the Improved U-Net models and functions.

• Good coding design. Well handwriting code.

• Sufficient docstring and comment in the code.

• Easy difficulty problem, no higher difficulty marks.

• Wrong pull request branch, from local:topic-recognition branch to shakes76:main branch.

[shakes76]

Marking

Good Practice (Design/Commenting, TF/Torch Usage)

Adequate design and implementation
Good spacing and comments
Header blocks missing -1

Recognition Problem

Solves problem
Driver Script present
File structure present
Shows Usage & Demo & Visualisation & Data usage
Module present
Commenting minimal -1
No Data leakage
Difficulty: Easy -10

Commit Log

Meaningful commit messages
Progressive commits used

Documentation

ReadMe OK, missing usage and other info -2
Model/technical explanation, could be better -1
Good Description and Comments
Markdown used and PDF submitted

Pull Request

Successful Pull Request (Working Algorithm Delivered on Time in InCorrect Branch) -2
Feedback required, remove model files and correct branch and remove gitignore files -2
Request Description OK

[shakes76]

Feedback marks possible +2 if the requested changes are made (see above).

[wangzhaomxy]

No feedback attempt and no feedback marks granted.