huggingface/text-generation-inference
5
0
Fork 0
mirror of https://github.com/huggingface/text-generation-inference.git synced 2025-09-12 12:54:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

test throughput

Browse Source
This commit is contained in:
Xuan Son Nguyen 2024-08-12 17:23:11 +02:00
parent b727e0aedc
commit d5d168a4d2
1 changed files with 24 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

42
load_tests/common.js
View File

@ -1,4 +1,4 @@
import { check } from 'k6';
import { check, sleep } from 'k6';
import { scenario } from 'k6/execution';
import http from 'k6/http';
import { Trend, Counter } from 'k6/metrics';
@ -33,13 +33,13 @@ export function get_options() {
// rate: 20,
// timeUnit: '1s',
// },
load_test: {
executor: 'constant-arrival-rate',
duration: '60s',
preAllocatedVUs: 10, // not enough RAM for 100 VUs
rate: 1,
timeUnit: '1s',
},
// load_test: {
// executor: 'constant-arrival-rate',
// duration: '60s',
// preAllocatedVUs: 10, // not enough RAM for 100 VUs
// rate: 1,
// timeUnit: '1s',
// },
// breakpoint: {
// executor: 'ramping-arrival-rate', //Assure load increase if the system slows
// preAllocatedVUs: 300,
@ -47,19 +47,21 @@ export function get_options() {
// { duration: '60s', target: 100 }, // just slowly ramp-up to a HUGE load
// ],
// },
// throughput: {
// executor: 'shared-iterations',
// vus: 100,
// iterations: 200,
// maxDuration: '40s',
// },
throughput: {
executor: 'shared-iterations',
vus: 100,
iterations: 500,
maxDuration: '400s',
},
},
};
}
function generate_payload(gpt, max_new_tokens) {
const input = gpt["conversations"][0]["value"];
return { "messages": [{ "role": "user", "content": input }], "temperature": 0, "model": `${model_id}`, "max_tokens": max_new_tokens }
return { "messages": [
{ "role": "user", "content": input.substring(0, 5000) }
], "temperature": 0, "model": `${model_id}`, "max_tokens": max_new_tokens }
}
export const options = get_options();
@ -71,12 +73,14 @@ export default function run() {
const res = http.post(`https://${host}/v1/chat/completions`, payload, {
headers,
});
if (res.status >= 400 && res.status < 500) {
return;
}
// if (res.status >= 400 && res.status < 500) {
// return;
// }
if (res.status !== 200) {
console.error(res.body);
console.error('status: ' + res.status);
return;
}
check(res, {
@ -94,4 +98,6 @@ export default function run() {
new_tokens.add(completion_tokens);
tokens.add(completion_tokens + prompt_tokens);
}
sleep(1);
}