Giter VIP home page Giter VIP logo

gpt3-tokenizer-java's Introduction

GPT3/4 Java Tokenizer

License: MIT GitHub Workflow Status Maven Central

This is a Java implementation of a GPT3/4 tokenizer, loosely ported from Tiktoken with the help of ChatGPT.

Usage Examples

Encoding Text to Tokens

GPT3Tokenizer tokenizer = new GPT3Tokenizer(Encoding.CL100K_BASE);
List<Integer> tokens = tokenizer.encode("example text here");

Decoding Tokens to Text

GPT3Tokenizer tokenizer = new GPT3Tokenizer(Encoding.CL100K_BASE);
List<Integer> tokens = Arrays.asList(123, 456, 789);
String text = tokenizer.decode(tokens);

Counting Number of Tokens in Chat Messages

var messages = List.of(
        new ChatMessage(ChatMessageRole.SYSTEM.value(), "You are a helpful assistant."),
        new ChatMessage(ChatMessageRole.USER.value(), "Hello there!")
);
var model = ModelType.GPT_3_5_TURBO;
var count = TokenCount.fromMessages(messages, model);
System.out.println("Prompt tokens: " + count);

Did you know...

  1. ...that all 3.5-turbo models released after 0613 now have tokenization counts for messages consistent with gpt-4 models?

  2. ...that OpenAI Tokenizer available at https://platform.openai.com/tokenizer uses p50k_base encoding, thus it doesn't count correctly tokens for gpt-3.5 and gpt-4 models? If you look for decent alternative, you may like: https://tiktokenizer.vercel.app/, but keep in mind that tokenization for messages of gpt-3.5 models released after 0613 was changed (see point above).

  3. ...that in cl100k_base encoding every sequence of up to 81 spaces is just a single token? So next time when someone tells you that passing YAML to ChatGPT is not efficient, you can argue that...

var tokenizer = ModelType.GPT_3_5_TURBO.getTokenizer();
var tokens = (List<Integer>) null;
for (var sb = new StringBuilder(" "); (tokens = tokenizer.encode(sb)).size() == 1; sb.append(' '))
    System.out.printf("`%s`'s token is %s, and that's %d space(s)!\n".replace("(s)", sb.length()==1?"":"s"), sb, tokens, sb.length());
`                                                                           `'s token is [14984], and that's 75 spaces!
`                                                                            `'s token is [56899], and that's 76 spaces!
`                                                                             `'s token is [59691], and that's 77 spaces!
`                                                                              `'s token is [82321], and that's 78 spaces!
`                                                                               `'s token is [40584], and that's 79 spaces!
`                                                                                `'s token is [98517], and that's 80 spaces!
`                                                                                 `'s token is [96529], and that's 81 spaces!

License

This project is licensed under the MIT License.

gpt3-tokenizer-java's People

Contributors

didalgolab avatar

Stargazers

 avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar  avatar

Watchers

 avatar  avatar  avatar  avatar

gpt3-tokenizer-java's Issues

gpt-4o support

it's throwing an error when new gpt-4o is informed as model.

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.