From 9c83e04f40f00a682fdca43a3f20a40f5adf9b12 Mon Sep 17 00:00:00 2001
From: Xuan Son Nguyen <son@huggingface.co>
Date: Wed, 14 Aug 2024 11:07:39 +0200
Subject: [PATCH] limit nb input tokens

---
 load_tests/common.js | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/load_tests/common.js b/load_tests/common.js
index 742064ca..aecaa0fd 100644
--- a/load_tests/common.js
+++ b/load_tests/common.js
@@ -50,7 +50,7 @@ export function get_options() {
             throughput: {
                 executor: 'shared-iterations',
                 vus: 16,
-                iterations: 200,
+                iterations: 500,
                 maxDuration: '400s',
             },
         },
@@ -58,7 +58,8 @@ export function get_options() {
 }
 
 function generate_payload(gpt, max_new_tokens) {
-    const input = gpt["conversations"][0]["value"];
+    const max_input_tokens = 10000 - max_new_tokens - 1;
+    const input = gpt["conversations"][0]["value"].substring(0, max_input_tokens);
     return {
         "prompt": `<|start_header_id|>system<|end_header_id|>\n\nYou are a helpful assistant<|eot_id|><|start_header_id|>user<|end_header_id|>\n\n${input}<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\n`,
         "max_tokens": max_new_tokens,