https://github.com/ComputerVisionLaboratory/cvlab_toolbox

CVLAB's tool box
https://github.com/ComputerVisionLaboratory/cvlab_toolbox

Last synced: 4 days ago
JSON representation

CVLAB's tool box

• Host: GitHub
• URL: https://github.com/ComputerVisionLaboratory/cvlab_toolbox
• Owner: ComputerVisionLaboratory
• Created: 2017-10-04T05:24:14.000Z (about 7 years ago)
• Default Branch: master
• Last Pushed: 2023-08-02T04:52:47.000Z (over 1 year ago)
• Last Synced: 2024-08-03T02:10:12.918Z (3 months ago)
• Language: Python
• Size: 18.2 MB
• Stars: 13
• Watchers: 4
• Forks: 8
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# cvlab_toolbox

This is the repository of CVLAB toolbox

## Usage

- Scikit-learn API

```python

import numpy as np

from numpy.random import randint, rand

from sklearn.model_selection import train_test_split

from sklearn.metrics import accuracy_score

from cvt.models import KernelMSM

dim = 100

n_class = 4

n_train, n_test = 20, 5

# input data X is list of vector sets (list of 2d-arrays)

X_train = [rand(randint(10, 20), dim) for i in range(n_train)]

X_test = [rand(randint(10, 20), dim) for i in range(n_test)]

# labels y is 1d-array

y_train = randint(0, n_class, n_train)

y_test = randint(0, n_class, n_test)

model = KernelMSM(n_subdims=3, sigma=0.01)

# fit

model.fit(X_train, y_train)

# predict

pred = model.predict(X_test)

print(accuracy_score(pred, y_test))

```

## Install

- pip

```bash

pip install -U git+https://github.com/ComputerVisionLaboratory/cvlab_toolbox

```

## Coding styles

- Follow `PEP8` as much as possible

  - [English](https://www.python.org/dev/peps/pep-0008/)

  - [日本語](http://pep8-ja.readthedocs.io/ja/latest/)

- Write a description as **docstring**

  ```python

  def PCA(X, whiten = False):

    '''

      apply PCA

      components, explained_variance = PCA(X)

      Parameters

      ----------

      X: ndarray, shape (n_samples, n_features)

        matrix of input vectors

      whiten: boolean

        if it is True, the data is treated as whitened

        on each dimensions (average is 0 and variance is 1)

      Returns

      -------

      components: ndarray, shape (n_features, n_features)

        the normalized component vectors

      explained_variance: ndarray, shape (n_features)

        the variance of each vectors

    '''

    ...

  ```

## Contribution rules

1. Make a pull request

2. Ask some lab members to review the code

3. when all agreements are taken, ask any admin member to merge it