tokenizer/tokenizer.hpp¶
Abstract tokenizer interface. More...
Namespaces¶
| Name |
|---|
| sgns |
| sgns::neoswarm |
| sgns::neoswarm::core |
Classes¶
| Name | |
|---|---|
| class | sgns::neoswarm::core::Tokenizer Abstract tokenizer interface. |
Detailed Description¶
Abstract tokenizer interface.
Date: 2026-05-06
Source code¶
#ifndef NEOSWARM_CORE_TOKENIZER_TOKENIZER_HPP
#define NEOSWARM_CORE_TOKENIZER_TOKENIZER_HPP
#include "common/error.hpp"
#include <memory>
#include <string>
#include <vector>
namespace sgns::neoswarm::core
{
class Tokenizer
{
public:
virtual ~Tokenizer() = default;
virtual outcome::result<std::vector<int>> Encode( const std::string& text ) const = 0;
virtual outcome::result<std::string> Decode( const std::vector<int>& ids ) const = 0;
virtual bool IsEOS( int token_id ) const = 0;
virtual int EosTokenId() const = 0;
virtual int BosTokenId() const = 0;
virtual size_t VocabSize() const = 0;
};
} // namespace sgns::neoswarm::core
#endif // NEOSWARM_CORE_TOKENIZER_TOKENIZER_HPP
Updated on 2026-07-25 at 22:56:57 +0000