llm

type DetokenizeRequest struct {
	Tokens []int `json:"tokens"`
}

type DetokenizeResponse struct {
	Content string `json:"content"`
}

type EmbeddingRequest struct {
	Content string `json:"content"`
}

type EmbeddingResponse struct {
	Embedding []float64 `json:"embedding"`
}

type GGML struct {
	// contains filtered or unexported fields
}

type LLM interface {
	Predict(context.Context, []int, string, func(api.GenerateResponse)) error
	Embedding(context.Context, string) ([]float64, error)
	Encode(context.Context, string) ([]int, error)
	Decode(context.Context, []int) (string, error)
	SetOptions(api.Options)
	Close()
	Ping(context.Context) error
}

type ModelRunner struct {
	Path        string // path to the model runner executable
	Accelerated bool
}

type Running struct {
	Port   int
	Cmd    *exec.Cmd
	Cancel context.CancelFunc

	*StatusWriter // captures error messages from the llama runner process
	// contains filtered or unexported fields
}

type StatusWriter struct {
	ErrCh      chan error
	LastErrMsg string
}

type TokenizeRequest struct {
	Content string `json:"content"`
}

type TokenizeResponse struct {
	Tokens []int `json:"tokens"`
}