Giter VIP home page Giter VIP logo

etlcdb's Introduction

Repack specification

  • The metadata will be stored in a JSON file.
  • Each record metadata will guarantee to contain at least the following keys.
{
    "title": "Record",
    "type": "object",
    "required": [
        "char",
        "dataset_source",
        "width",
        "height",
        "seek_start",
        "seek_end",
    ],
    "properties": {
        "char": {
            "type": "string",
            "description": "The label char. If the label is invalid, it will start with 'null_'.",
        },
        "dataset_source": {
            "type": "string",
            "description": "The etlcdb data source file name.",
        },
        "width": {
            "type": "number",
            "description": "The image width.",
        },
        "height": {
            "type": "number",
            "description": "The image height.",
        },
        "seek_start": {
            "type": "number",
            "description": "The PNG image data (in bytes) starting position in the packed image data binary which contains image data for mulitple records.",
        },
        "seek_end": {
            "type": "number",
            "description": "The PNG image data (in bytes) ending position in the packed image data binary which contains image data for mulitple records.",
        },
    },
}
  • There will be other properties in each records based on the source dataset.
  • The PNG images can be either in grayscale (pixel values range from 0 to 255) or in binary (pixel values are either 0 or 1). Because of that caution should be taken when normalizing data in range of [0,1] before passing it to the machine learning model.

Cleaning dataset status

  • ETL2
  • ETL3
  • ETL4
  • ETL5
  • ETL6
  • ETL7
  • ETL8G
  • ETL8B2
  • ETL9G
  • ETL9B

ETL2

  • 52786 samples
  • 2168 classes (symbols, hiragana, katakana, kanji, english alphabet)
  • image_width: 60px
  • image_height: 60px

ETL3

This dataset contains numbers (0-9), English alphabet letters and symbols with each has 200 samples.

ETL4

This dataset contains 51 hiragana characters with 120 samples each.

ETL5

This dataset contains 51 katakana characters with 208 samples each.

ETL8B

  • 153916 samples
  • 956 classes (hiragana and kanji)
  • 161 samples each class
  • image_width: 64px
  • image_height: 63px

ETL9G

  • 607200 samples
  • 3036 classes (hiragana and kanji)
  • 200 samples each class
  • record_length: 8199 bytes
  • image_width: 128px
  • image_height: 127px

etlcdb's People

Contributors

ichisadashioko avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar

Forkers

phanminhtuananh

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    ๐Ÿ–– Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. ๐Ÿ“Š๐Ÿ“ˆ๐ŸŽ‰

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google โค๏ธ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.