34 lines
1.1 KiB
TypeScript
34 lines
1.1 KiB
TypeScript
import { LargeLanguageModel } from "@/lib/schemas";
|
|
import { readSettings } from "../../main/settings";
|
|
import { Message } from "../ipc_types";
|
|
|
|
import { findLanguageModel } from "./findLanguageModel";
|
|
|
|
// Estimate tokens (4 characters per token)
|
|
export const estimateTokens = (text: string): number => {
|
|
return Math.ceil(text.length / 4);
|
|
};
|
|
|
|
export const estimateMessagesTokens = (messages: Message[]): number => {
|
|
return messages.reduce(
|
|
(acc, message) => acc + estimateTokens(message.content),
|
|
0,
|
|
);
|
|
};
|
|
|
|
const DEFAULT_CONTEXT_WINDOW = 128_000;
|
|
|
|
export async function getContextWindow() {
|
|
const settings = readSettings();
|
|
const modelOption = await findLanguageModel(settings.selectedModel);
|
|
return modelOption?.contextWindow || DEFAULT_CONTEXT_WINDOW;
|
|
}
|
|
|
|
// Most models support at least 8000 output tokens so we use it as a default value.
|
|
const DEFAULT_MAX_TOKENS = 8_000;
|
|
|
|
export async function getMaxTokens(model: LargeLanguageModel) {
|
|
const modelOption = await findLanguageModel(model);
|
|
return modelOption?.maxOutputTokens || DEFAULT_MAX_TOKENS;
|
|
}
|