sentencepiece

zechenzhangAGI's avatarfrom zechenzhangAGI

Language-independent tokenizer treating text as raw Unicode. Supports BPE and Unigram algorithms. Fast (50k sentences/sec), lightweight (6MB memory), deterministic vocabulary. Used by T5, ALBERT, XLNet, mBART. Train on raw text without pre-tokenization. Use when you need multilingual support, CJK languages, or reproducible tokenization.

⭐62stars🔀2forks📁View on GitHub🕐Updated Nov 16, 2025
[Developer Tools]

When & Why to Use This Skill

SentencePiece is a high-performance, language-independent subword tokenizer designed for neural network-based text processing. By treating text as raw Unicode strings, it eliminates the need for language-specific pre-tokenization, making it an essential tool for building multilingual NLP models. It supports both BPE and Unigram algorithms and is the core tokenization technology behind major models like T5, ALBERT, and XLNet.

Use Cases

  • Multilingual Model Development: Creating unified vocabularies for AI models that support dozens of languages simultaneously without complex preprocessing rules.
  • CJK Language Processing: Efficiently handling Chinese, Japanese, and Korean text where traditional whitespace-based tokenization is not applicable.
  • High-Throughput Production Pipelines: Leveraging its 50,000 sentences/second speed and 6MB memory footprint for low-latency inference in resource-constrained environments.
  • Data Augmentation for Robustness: Using subword regularization (sampling different tokenizations) to improve the robustness and accuracy of machine translation and text generation models.
namesentencepiece
descriptionLanguage-independent tokenizer treating text as raw Unicode. Supports BPE and Unigram algorithms. Fast (50k sentences/sec), lightweight (6MB memory), deterministic vocabulary. Used by T5, ALBERT, XLNet, mBART. Train on raw text without pre-tokenization. Use when you need multilingual support, CJK languages, or reproducible tokenization.
version1.0.0
authorOrchestra Research
licenseMIT
tags[Tokenization, SentencePiece, Language-Independent, BPE, Unigram, Multilingual, CJK Languages, Unicode, Deterministic, Google]
dependencies[sentencepiece, transformers]

SentencePiece - Language-Independent Tokenization

Unsupervised tokenizer that works on raw text without language-specific preprocessing.

When to use SentencePiece

Use SentencePiece when:

  • Building multilingual models (no language-specific rules)
  • Working with CJK languages (Chinese, Japanese, Korean)
  • Need reproducible tokenization (deterministic vocabulary)
  • Want to train on raw text (no pre-tokenization needed)
  • Require lightweight deployment (6MB memory, 50k sentences/sec)

Performance:

  • Speed: 50,000 sentences/sec
  • Memory: ~6MB for loaded model
  • Languages: All (language-independent)

Use alternatives instead:

  • HuggingFace Tokenizers: Faster training, more flexibility
  • tiktoken: OpenAI models (GPT-3.5/4)
  • BERT WordPiece: English-centric tasks

Quick start

Installation

# Python
pip install sentencepiece

# C++ (requires CMake)
git clone https://github.com/google/sentencepiece.git
cd sentencepiece
mkdir build && cd build
cmake .. && make -j $(nproc)
sudo make install

Train model

# Command-line (BPE with 8000 vocab)
spm_train --input=data.txt --model_prefix=m --vocab_size=8000 --model_type=bpe

# Python API
import sentencepiece as spm

spm.SentencePieceTrainer.train(
    input='data.txt',
    model_prefix='m',
    vocab_size=8000,
    model_type='bpe'
)

Training time: ~1-2 minutes for 100MB corpus

Encode and decode

import sentencepiece as spm

# Load model
sp = spm.SentencePieceProcessor(model_file='m.model')

# Encode to pieces
pieces = sp.encode('This is a test', out_type=str)
print(pieces)  # ['▁This', '▁is', '▁a', '▁test']

# Encode to IDs
ids = sp.encode('This is a test', out_type=int)
print(ids)  # [284, 47, 11, 1243]

# Decode
text = sp.decode(ids)
print(text)  # "This is a test"

Language-independent design

Whitespace as symbol (▁)

text = "Hello world"
pieces = sp.encode(text, out_type=str)
print(pieces)  # ['▁Hello', '▁world']

# Decode preserves spaces
decoded = sp.decode_pieces(pieces)
print(decoded)  # "Hello world"

Key principle: Treat text as raw Unicode, whitespace = ▁ (meta symbol)

Tokenization algorithms

BPE (Byte-Pair Encoding)

spm.SentencePieceTrainer.train(
    input='data.txt',
    model_prefix='bpe_model',
    vocab_size=16000,
    model_type='bpe'
)

Used by: mBART

Unigram (default)

spm.SentencePieceTrainer.train(
    input='data.txt',
    model_prefix='unigram_model',
    vocab_size=8000,
    model_type='unigram'
)

Used by: T5, ALBERT, XLNet

Training configuration

Essential parameters

spm.SentencePieceTrainer.train(
    input='corpus.txt',
    model_prefix='m',
    vocab_size=32000,
    model_type='unigram',
    character_coverage=0.9995,  # 1.0 for CJK
    user_defined_symbols=['[SEP]', '[CLS]'],
    unk_piece='<unk>',
    num_threads=16
)

Character coverage

Language Type Coverage Rationale
English 0.9995 Most common chars
CJK (Chinese) 1.0 All characters needed
Multilingual 0.9995 Balance

Encoding options

Subword regularization

# Sample different tokenizations
for _ in range(3):
    pieces = sp.encode('tokenization', out_type=str, enable_sampling=True, alpha=0.1)
    print(pieces)

# Output (different each time):
# ['▁token', 'ization']
# ['▁tok', 'en', 'ization']

Use case: Data augmentation for robustness.

Common patterns

T5-style training

spm.SentencePieceTrainer.train(
    input='c4_corpus.txt',
    model_prefix='t5',
    vocab_size=32000,
    model_type='unigram',
    user_defined_symbols=[f'<extra_id_{i}>' for i in range(100)],
    unk_id=2,
    eos_id=1,
    pad_id=0
)

Integration with transformers

from transformers import T5Tokenizer

# T5 uses SentencePiece internally
tokenizer = T5Tokenizer.from_pretrained('t5-base')
inputs = tokenizer('translate English to French: Hello', return_tensors='pt')

Performance benchmarks

Training speed

Corpus BPE (16k) Unigram (8k)
100 MB 1-2 min 3-4 min
1 GB 10-15 min 30-40 min

Tokenization speed

  • SentencePiece: 50,000 sentences/sec
  • HF Tokenizers: 200,000 sentences/sec (4× faster)

Supported models

T5 family: t5-base, t5-large (32k vocab, Unigram) ALBERT: albert-base-v2 (30k vocab, Unigram) XLNet: xlnet-base-cased (32k vocab, Unigram) mBART: facebook/mbart-large-50 (250k vocab, BPE)

References

Resources