/Noise-Adaption-Layer

This is the Pytorch implementation of paper--Training deep neural-networks using a noise adaptation layer.

Primary LanguagePythonMIT LicenseMIT

Noise-Adaption-Layer

This is the Pytorch implementation of paper--Training deep neural-networks using a noise adaptation layer, but not official.

Requirements

Python 3.7
Pytorch 1.8.0
pandas
numpy

Dataset

There are three datasets and I perturb the noise in data_utils.py file.If your are interested in perturbing noise in dataset, you could check noise.py.

And the datasets I prerprocessed can be downloaded here with key: sn86. (BaiduNetDisk)

How to run it?

You could check the shell script and you will get to know how it could work.