dmlc/ experimental-lda

C++

Shell

Need help with experimental-lda?

Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

dmlc

126 Stars

63 Forks

Other

91 Commits

2 Opened issues

Services available

Need anything else?

Contributors list

Readme

Single Machine implementation of LDA

Name: experimental-lda
Brand: experimental-lda
SKU: //dmlc/experimental-lda
Rating: 2.065 (126 reviews)

Modules

```
parallelLDA
```
contains various implementation of multi threaded LDA
```
singleLDA
```
contains various implementation of single threaded LDA
```
topwords
```
a tool to explore topics learnt by the LDA/HDP
```
perplexity
```
a tool to calculate perplexity on another dataset using word|topic matrix
```
datagen
```
packages txt files for our program
```
preprocessing
```
for converting from UCI or cLDA to simple txt file having one document per line

Organisation

All codes are under
```
src
```
within respective folder
For running Topic Models many template scripts are provided under
```
scripts
```
```
data
```
is a placeholder folder where to put the data
```
build
```
and
```
dist
```
folder will be created to hold the executables

Requirements

gcc >= 5.0 or Intel® C++ Compiler 2016 for using C++14 features
split >= 8.21 (part of GNU coreutils)

How to use

We will show how to run our LDA on an UCI bag of words dataset

First of all compile by hitting make

     make

Download example dataset from UCI repository. For this a script has been provided.

     scripts/get_data.sh

Prepare the data for our program

     scripts/prepare.sh data/nytimes 1

For other datasets replace nytimes with dataset name or location.

Run LDA!

     scripts/lda_runner.sh

Inside the

lda_runner.sh

all the parameters e.g. number of topics, hyperparameters of the LDA, number of threads etc. can be specified. By default the outputs are stored under

out/

. Also you can specify which inference algorithm of LDA you want to run: 1.

simpleLDA

: Plain vanilla Gibbs sampling by Griffiths04 2.

sparseLDA

: Sparse LDA of Yao09 3.

aliasLDA

: Alias LDA 4.

FTreeLDA

: F++LDA (inspired from Yu14 5.

lightLDA

: light LDA of Yuan14

The make file has some useful features:

if you have Intel® C++ Compiler, then you can instead

     make intel

or if you want to use Intel® C++ Compiler's cross-file optimization (ipo), then hit

     make inteltogether

Also you can selectively compile individual modules by specifying

     make

or clean individually by

     make clean-

Performance

Based on our evaluation F++LDA works the best in terms of both speed and perplexity on a held-out dataset. For example on Amazon EC2 c4.8xlarge, we obtained more than 25 million/tokens per second. Below we provide performance comparison against various inference procedures on publicaly available datasets.

Datasets

| Dataset | V | L | D | L/V | L/D | | ------------ | --------: | --------------: | -----------: | --------: | --------: | | NY Times | 101,330 | 99,542,127 | 299,753 | 982.36 | 332.08 | | PubMed | 141,043 | 737,869,085 | 8,200,000 | 5,231.52 | 89.98 | | Wikipedia | 210,218 | 1,614,349,889 | 3,731,325 | 7,679.41 | 432.65 |

Experimental datasets and their statistics.

denotes vocabulary size,

denotes the number of training tokens,

denotes the number of documents,

L/V

indicates the average number of occurrences of a word,

L/D

indicates the average length of a document.