regsets 0.1.1

A collection of regression datasets, featuring PyTorch-like dataset classes.

🎛️ Regression Datasets

📋 Table of Contents

1. Usage
2. Datasets
3. License
4. Contact
5. Acknowledgments

This repository contains a collection of various regression datasets. I have unified their data format to make them easier to read and process. Additionally, I have included code that follows the PyTorch Datasets structure, which allows users to automatically download and load the datasets. All datasets come with a permissive license, permitting their use for research purposes.

1. Usage

This repository contains datasets for Vision, Audio, and Text. The corresponding files are located in the following folders:

• ./vision
• ./audio
• ./text

Each folder contains multiple [dataset].py files. You can directly import these files to automatically download and load the datasets. Alternatively, you can copy the desired [dataset].py file into your project to load the dataset locally.

📸 Example Usage of Vision Datasets

from vision.utkface import UTKFace

utkface_trainset = UTKFace(root="./data", split="train", download=True)

for image, label in utkface_trainset:
    ...

🎧 Example Usage of Audio Datasets

from audio.vcc2018 import VCC2018

vcc2018_trainset = VCC2018(root="./data", split="train", download=True)

for audio, sample_rate, label in vcc2018_trainset:
    ...

📝 Example Usage of Text Datasets

from text.amazon_review import Amazon_Review

amazon_review_trainset = Amazon_Review(root="./data", split="train", download=True)

for texts, label in amazon_review_trainset:
    (ori, aug_0, aug_1) = texts
    ...

(back to top)

2. Datasets

For datasets that do not provide a predefined train-test split, I randomly sample 80% of the data for training and reserve the remaining 20% for testing. Details for each dataset are provided below.

📸 Vision

Dataset	# Training Data	# Dev Data	# Test Data	Target Range
UTKFace	18,964	-	4,741	[1, 116]

🎧 Audio

Dataset	# Training Data	# Dev Data	# Test Data	Target Range
BVCC	4,974	1,066	1,066	[1, 5]
VCC2018	16,464	-	4,116	[1, 5]

📝 Text

Dataset	# Training Data	# Dev Data	# Test Data	Target Range
Amazon Review	250,000	25,000	650,000	[0, 4]
Yelp Review	250,000	25,000	50,000	[0, 4]

(back to top)

3. License

Distributed under the MIT License. See LICENSE for more information.

(back to top)

4. Contact

• Pin-Yen Huang ([e-mail])

(back to top)

5. Acknowledgments

• PyTorch
• UTKFace
• VCC2018
• BVCC
• USB
• Amazon Review
• Yelp Review
• README Template

(back to top)