From 0a888733ce3d3eef5a38cfe10e697290e9f021d7 Mon Sep 17 00:00:00 2001
From: kuangquanshuisn <kuangquanshuisn@gmail.com>
Date: Mon, 28 Oct 2024 11:28:04 +0800
Subject: [PATCH] =?UTF-8?q?=F0=9F=90=9B=20fix:=20Resolve=20the=20issue=20o?=
 =?UTF-8?q?f=20the=20Huggingface=20API=20interrupting=20when=20the=20outpu?=
 =?UTF-8?q?t=20exceeds=20140=20tokens?=
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

解决Huggingface API在输出超过140多个token时中断的问题
---
 src/libs/agent-runtime/huggingface/index.ts | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/src/libs/agent-runtime/huggingface/index.ts b/src/libs/agent-runtime/huggingface/index.ts
index 78aa41b36fabc..b31a12b1af276 100644
--- a/src/libs/agent-runtime/huggingface/index.ts
+++ b/src/libs/agent-runtime/huggingface/index.ts
@@ -21,8 +21,10 @@ export const LobeHuggingFaceAI = LobeOpenAICompatibleFactory({
   },
   customClient: {
     createChatCompletionStream: (client: HfInference, payload, instance) => {
+      const { max_tokens = 4096} = payload;
       const hfRes = client.chatCompletionStream({
         endpointUrl: instance.baseURL,
+        max_tokens: max_tokens,
         messages: payload.messages,
         model: payload.model,
         stream: true,