metadata
dataset_info:
features:
- name: image
dtype: image
- name: 'Unnamed: 0'
dtype: int64
- name: articleType
dtype: string
- name: productDisplayName
dtype: string
- name: articleType_label
dtype: int64
splits:
- name: train
num_bytes: 140935233.8
num_examples: 9300
download_size: 122008451
dataset_size: 140935233.8
Dataset Card for "fashion_data"
The dataset consists of 31 classes and each class has 300 images along with the productDisplayName. This was created from the a larger dataset which you can find here.
The purpose of creating this dataset was to make a Image search engine from a database where the input is either an image, text or audio and the output is a set of images similar to the input.
You can find the project implementation here.