From ea6a5b03e077526896071da80530bebb94eb390b Mon Sep 17 00:00:00 2001
From: Jan Pawellek <2392542+janpawellek@users.noreply.github.com>
Date: Mon, 19 Jun 2023 02:01:15 +0200
Subject: [PATCH] Fix output final text for HuggingFaceTextGenInference when
 streaming (#6211)

The LLM integration
[HuggingFaceTextGenInference](https://github.com/hwchase17/langchain/blob/master/langchain/llms/huggingface_text_gen_inference.py)
already has streaming support.

However, when streaming is enabled, it always returns an empty string as
the final output text when the LLM is finished. This is because `text`
is instantiated with an empty string and never updated.

This PR fixes the collection of the final output text by concatenating
new tokens.
---
 langchain/llms/huggingface_text_gen_inference.py | 1 +
 1 file changed, 1 insertion(+)

diff --git a/langchain/llms/huggingface_text_gen_inference.py b/langchain/llms/huggingface_text_gen_inference.py
index 3d17c73418..cab1178983 100644
--- a/langchain/llms/huggingface_text_gen_inference.py
+++ b/langchain/llms/huggingface_text_gen_inference.py
@@ -169,4 +169,5 @@ class HuggingFaceTextGenInference(LLM):
                 if not token.special:
                     if text_callback:
                         text_callback(token.text)
+                    text += token.text
         return text