sphinx-quickstart on Wed Aug 14 16:39:23 2019. You can adapt this file completely to your liking, but it should at least contain the root toctree directive.

github_url

https://github.com/p768lwy3/torecsys

Welcome to torecsys’s documentation!

Recommendation System in PyTorch.

This package is an implementation of several famous recommendation systelm algorithm in PyTorch, including click-through-rate prediction, learning-to-ranking and embedding. ToR[e]csys also comes with data loader, layers-level implementation, self-defined architecture of models. It’s open-source software, released under the MIT license.

Minimial Requirements

  • Numpy >= 1.17.0

  • Pandas >= 0.24.2

  • PyTorch >= 1.2

Installation

Install with pip:

pip install torecsys

Install with source code:

git clone https://github.com/p768lwy3/torecsys.git
cd ./torecsys
python setup.py build
python setup.py install

Notations in documentation

notations

Notation

Refer to

T

torch.Tensor

B

batch size

E

embedding size

H_i

output size of i-th hidden layer

N

number of fields

I

input sizes of any layers required inputs_size

O

output size of any layers requried outputs_size

V

total number of words in a vocabulary set

S

total number of samples, e.g. negative samples

API documentation

Indices and tables