Skip to main content

AI21 Labs Tokenizer

A SentencePiece based tokenizer for production uses with AI21's models

Test Package version Supported Python versions Poetry Supported Python versions License


Prerequisites

  • If you wish to use the tokenizers for Jamba Mini or Jamba Large, you will need to request access to the relevant model's HuggingFace repo:

Installation

pip

pip install ai21-tokenizer

poetry

poetry add ai21-tokenizer

Usage

Basic Usage

from ai21_tokenizer import Tokenizer

# Create tokenizer (defaults to Jamba Mini)
tokenizer = Tokenizer.get_tokenizer()

# Encode text to token IDs
text = "Hello, world!"
encoded = tokenizer.encode(text)
print(f"Encoded: {encoded}")

# Decode token IDs back to text
decoded = tokenizer.decode(encoded)
print(f"Decoded: {decoded}")

Specific Tokenizer Selection

from ai21_tokenizer import Tokenizer, PreTrainedTokenizers

# Jamba Mini tokenizer
tokenizer = Tokenizer.get_tokenizer(PreTrainedTokenizers.JAMBA_MINI_TOKENIZER)

# Jamba Large tokenizer
tokenizer = Tokenizer.get_tokenizer(PreTrainedTokenizers.JAMBA_LARGE_TOKENIZER)

Async Usage

import asyncio
from ai21_tokenizer import Tokenizer

async def main():
    tokenizer = await Tokenizer.get_async_tokenizer()

    text = "Hello, world!"
    encoded = await tokenizer.encode(text)
    decoded = await tokenizer.decode(encoded)

    print(f"Original: {text}")
    print(f"Encoded: {encoded}")
    print(f"Decoded: {decoded}")

asyncio.run(main())

Advanced Token Operations

# Convert between tokens and IDs
tokens = tokenizer.convert_ids_to_tokens(encoded)
print(f"Tokens: {tokens}")

ids = tokenizer.convert_tokens_to_ids(tokens)
print(f"IDs: {ids}")

Direct Class Usage

from ai21_tokenizer import SyncJambaTokenizer

# Using local model file
model_path = "/path/to/your/tokenizer.model"
tokenizer = SyncJambaTokenizer(model_path=model_path)

text = "Hello, world!"
encoded = tokenizer.encode(text)
decoded = tokenizer.decode(encoded)

Async Direct Class Usage

from ai21_tokenizer import AsyncJambaTokenizer

async def main():
    model_path = "/path/to/your/tokenizer.model"
    tokenizer = await AsyncJambaTokenizer.create(model_path=model_path)

    text = "Hello, world!"
    encoded = await tokenizer.encode(text)
    decoded = await tokenizer.decode(encoded)

asyncio.run(main())

For more examples, please see our examples folder.

Metadata

Release files for ai21-tokenizer 1.1.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for ai21-tokenizer 1.1.0
File Size Uploaded Provenance Actions
ai21_tokenizer-1.1.0.tar.gz 2.6 MB Without provenance Details for ai21_tokenizer-1.1.0.tar.gz

Built distribution (wheel)

Table of built distributions (wheels) for ai21-tokenizer 1.1.0
File Interpreter ABI Platform Actions
ai21_tokenizer-1.1.0-py3-none-any.whl Python 3 none any Details for ai21_tokenizer-1.1.0-py3-none-any.whl

Total release size: 5.3 MB

Release files / ai21_tokenizer-1.1.0.tar.gz

Download URL ai21_tokenizer-1.1.0.tar.gz
Size 2.6 MB
Tags Source
SHA-256 checksum
How to use checksums
aa511dc7d9717176fe37b24868cef5f8a568c4d3ab6ed336da69da10d7771c92
BLAKE2b-256 checksum
How to use checksums
0a96275e0000aaad03fe118afa964e669d69e0a8706dfbb079fde6837bf602c1
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.9.23

Release files / ai21_tokenizer-1.1.0-py3-none-any.whl

Download URL ai21_tokenizer-1.1.0-py3-none-any.whl
Size 2.7 MB
Tags Python 3
SHA-256 checksum
How to use checksums
9402c8387c09a22b13995ea485b6d3ebe119c380bfaae733f2541da73457f7ed
BLAKE2b-256 checksum
How to use checksums
13fbd7ff8467745892a6fe65b0be7f6679654c9c6d1aba793bf6ccb8198120a5
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
No
Uploaded via twine/6.1.0 CPython/3.9.23

Security for ai21-tokenizer 1.1.0

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

No provenance found

None of the files in this release were uploaded with provenance metadata.

A note on package security

PyPI is a repository for distributing packages, not a service for auditing them. While third-party security researchers and our staff do their best to detect and remove malicious packages, PyPI cannot guarantee the quality, security or suitability of any given package.

To learn more about protecting yourself from malicious actors, visit our package security documentation.

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page