llame

type CompletionQuery struct {
	Prompt      string  `json:"prompt"`
	Temperature float32 `json:"float,omitempty"`
	NPredict    int     `json:"n_predict,omitempty"` // Default: `-1`, where `-1` is infinity. Set the maximum number of tokens to predict when generating text.
	Stream      bool    `json:"stream,omitempty"`
}

type GitFiles struct {
	Tracked   []string
	Untracked []string
}

type LlamaModel struct {
	RequestTimeout time.Duration
	// contains filtered or unexported fields
}

type PromptFormat struct {
	Template        string `json:"template"`
	HistoryTemplate string `json:"historyTemplate"`
	Char            string `json:"char"`
	CharMsgPrefix   string `json:"charMsgPrefix"`
	CharMsgSuffix   string `json:"charMsgSuffix"`
	User            string `json:"user"`
	UserMsgPrefix   string `json:"userMsgPrefix"`
	UserMsgSuffix   string `json:"userMsgSuffix"`
	Stops           string `json:"stops"`
}

type PromptFormats map[string]PromptFormat

type StreamData struct {
	Content    string `json:"content"`    // The content being streamed
	Stop       bool   `json:"stop"`       // Indicates whether the stream should stop
	IdSlot     int    `json:"id_slot"`    // Slot to which the task is assigned
	Multimodal bool   `json:"multimodal"` // Indicates if the response is multimodal
	Index      int    `json:"index"`      // Index of the response in the stream
}

type StreamResponse struct {
	StreamData
	Error error
}

type TextMessage struct {
	Name    string // User or Assistant
	Message string
}