mirror of
https://github.com/ggerganov/llama.cpp.git
synced 2024-12-27 22:59:24 +01:00
7ee76e45af
* expose simple web interface on root domain * embed index and add --path for choosing static dir * allow server to multithread because web browsers send a lot of garbage requests we want the server to multithread when serving 404s for favicon's etc. To avoid blowing up llama we just take a mutex when it's invoked. * let's try this with the xxd tool instead and see if msvc is happier with that * enable server in Makefiles * add /completion.js file to make it easy to use the server from js * slightly nicer css * rework state management into session, expose historyTemplate to settings --------- Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
82 lines
2.2 KiB
JavaScript
82 lines
2.2 KiB
JavaScript
const paramDefaults = {
|
|
stream: true,
|
|
n_predict: 500,
|
|
temperature: 0.2,
|
|
stop: ["</s>"]
|
|
};
|
|
|
|
/**
|
|
* This function completes the input text using a llama dictionary.
|
|
* @param {object} params - The parameters for the completion request.
|
|
* @param {object} controller - an instance of AbortController if you need one, or null.
|
|
* @param {function} callback - The callback function to call when the completion is done.
|
|
* @returns {string} the completed text as a string. Ideally ignored, and you get at it via the callback.
|
|
*/
|
|
export const llamaComplete = async (params, controller, callback) => {
|
|
if (!controller) {
|
|
controller = new AbortController();
|
|
}
|
|
const completionParams = { ...paramDefaults, ...params };
|
|
|
|
// we use fetch directly here becasue the built in fetchEventSource does not support POST
|
|
const response = await fetch("/completion", {
|
|
method: 'POST',
|
|
body: JSON.stringify(completionParams),
|
|
headers: {
|
|
'Connection': 'keep-alive',
|
|
'Content-Type': 'application/json',
|
|
'Accept': 'text/event-stream'
|
|
},
|
|
signal: controller.signal,
|
|
});
|
|
|
|
const reader = response.body.getReader();
|
|
const decoder = new TextDecoder();
|
|
|
|
let content = "";
|
|
|
|
try {
|
|
|
|
let cont = true;
|
|
|
|
while (cont) {
|
|
const result = await reader.read();
|
|
if (result.done) {
|
|
break;
|
|
}
|
|
|
|
// sse answers in the form multiple lines of: value\n with data always present as a key. in our case we
|
|
// mainly care about the data: key here, which we expect as json
|
|
const text = decoder.decode(result.value);
|
|
|
|
// parse all sse events and add them to result
|
|
const regex = /^(\S+):\s(.*)$/gm;
|
|
for (const match of text.matchAll(regex)) {
|
|
result[match[1]] = match[2]
|
|
}
|
|
|
|
// since we know this is llama.cpp, let's just decode the json in data
|
|
result.data = JSON.parse(result.data);
|
|
content += result.data.content;
|
|
|
|
// callack
|
|
if (callback) {
|
|
cont = callback(result) != false;
|
|
}
|
|
|
|
// if we got a stop token from server, we will break here
|
|
if (result.data.stop) {
|
|
break;
|
|
}
|
|
}
|
|
} catch (e) {
|
|
console.error("llama error: ", e);
|
|
throw e;
|
|
}
|
|
finally {
|
|
controller.abort();
|
|
}
|
|
|
|
return content;
|
|
}
|