search

ankitdhall / learning_embeddings

Code for CVPR-W 2020 paper "Hierarchical Image Classification using Entailment Cone Embeddings" https://arxiv.org/abs/2004.03459
https://ankitdhall.github.io/project/learning-representations-for-images-with-hierarchical-labels/
21 stars 7 forks source link

About the usage of CIFAR10 dataset in the code repo #33

Open zyh10655 opened 1 year ago

zyh10655 commented 1 year ago

Greetings, I notice in the ethec_experiments.py of the Adam1X branch, both ETHECExperiment() and ETHEC2D() need a CIFAR10 object. May I ask what is CIFAR10 used here? If the aiming dataset is ETHEC1.0, why do we still need to define the label map of CIFAR10? Sincerely thanks for any possible response.

ankitdhall commented 1 year ago

Do you mean this line? https://github.com/ankitdhall/learning_embeddings/blob/Adam1x/network/ethec_experiments.py#L48

If that is the case, it was to reuse the experimental setup created for the CIFAR10 dataset. The dataset-specific details are overridden in the child class i.e. ETHEC2D for example. In hindsight, I could have designed it better. I hope this answers your query.