mirror of
https://github.com/nomic-ai/gpt4all.git
synced 2025-12-22 03:24:34 +00:00
ee3469ba6c
Improves output quality by making these tokenizers more closely match the behavior of the huggingface `tokenizers` based BPE tokenizers these models were trained with. Featuring: * Fixed unicode handling (via ICU) * Fixed BPE token merge handling * Complete added vocabulary handling
40 lines
1.1 KiB
C++
40 lines
1.1 KiB
C++
#ifndef GPTJ_H
|
|
#define GPTJ_H
|
|
|
|
#include <string>
|
|
#include <functional>
|
|
#include <vector>
|
|
#include "llmodel.h"
|
|
#include "tokenizer/bpe.h"
|
|
|
|
class GPTJPrivate;
|
|
class GPTJ : public LLModel {
|
|
public:
|
|
GPTJ();
|
|
~GPTJ();
|
|
|
|
bool loadModel(const std::string &modelPath) override;
|
|
bool isModelLoaded() const override;
|
|
size_t stateSize() const override;
|
|
size_t saveState(uint8_t *dest) const override;
|
|
size_t restoreState(const uint8_t *src) override;
|
|
void prompt(const std::string &prompt,
|
|
std::function<bool(int32_t)> promptCallback,
|
|
std::function<bool(int32_t, const std::string&)> responseCallback,
|
|
std::function<bool(bool)> recalculateCallback,
|
|
PromptContext &ctx) override;
|
|
void setThreadCount(int32_t n_threads) override;
|
|
int32_t threadCount() const override;
|
|
|
|
protected:
|
|
void recalculateContext(PromptContext &promptCtx,
|
|
std::function<bool(bool)> recalculate) override;
|
|
|
|
private:
|
|
GPTJPrivate *d_ptr;
|
|
std::unique_ptr<bpecpp::AdditionalVocabAdapter> m_tokav;
|
|
std::unique_ptr<bpecpp::BPE> m_bpe;
|
|
};
|
|
|
|
#endif // GPTJ_H
|