Skip to main content

Crate llm_tokenizer

Crate llm_tokenizer 

Source

Re-exports§

pub use cache::CacheConfig;
pub use cache::CacheStats;
pub use cache::CachedTokenizer;
pub use cache::L0Cache;
pub use cache::L1Cache;
pub use cache::TokenizerFingerprint;
pub use chat_template::ChatTemplateState;
pub use factory::create_tokenizer;
pub use factory::create_tokenizer_from_file;
pub use factory::create_tokenizer_with_chat_template;
pub use factory::TokenizerType;
pub use huggingface::HuggingFaceTokenizer;
pub use mock::MockTokenizer;
pub use registry::LoadError;
pub use registry::LoadOutcome;
pub use registry::TokenizerRegistry;
pub use sequence::Sequence;
pub use stop::SequenceDecoderOutput;
pub use stop::StopSequenceConfig;
pub use stop::StopSequenceDecoder;
pub use stream::DecodeStream;
pub use tiktoken::TiktokenModel;
pub use tiktoken::TiktokenTokenizer;
pub use traits::Decoder;
pub use traits::Encoder;
pub use traits::Encoding;
pub use traits::SpecialTokens;
pub use traits::TokenIdType;
pub use traits::Tokenizer as TokenizerTrait;

Modules§

cache
Tokenizer Caching Layer
chat_template
Chat template support for tokenizers using Jinja2 templates
factory
hub
huggingface
mock
Mock tokenizer implementation for testing
registry
Tokenizer Registry for dynamic tokenizer loading
sequence
stop
stream
tiktoken
traits

Structs§

Tokenizer
Main tokenizer wrapper that provides a unified interface for different tokenizer implementations