This repository contains the implementation of the Deep Learning course assignment at the University of Trento.
Given the person re-identification dataset Market1501, the assignment is composed by two task
Task 1 - Classification Task: The dataset contains a csv file with attributes annotations. Each image of the training set is labeled with 27 attributes that includes age, hair length, gender, length of lower/upper-body clothing, wearing hat, anong others, more details here. The task consists on predicting such attributes.
Task 2 - Person Re-Identification Task: This is an image retrieval task, that returns a list of images that corresponds to the same identity of a given query image. The main framework used in this implementation is PyTorch.
To execute the code in this repo:
- Go to http://colab.research.google.com/
- Upload the "DL_Project_wandb_embedding.ipynb" notebook or select the github tab and paste deayalar/deeplearning_unitn in the url (use the branch triplet)
- Run all cells in the Notebook