From d9b0f2c9c3a66deab83599670f8565e16d14b3f1 Mon Sep 17 00:00:00 2001
From: oobabooga <112222186+oobabooga@users.noreply.github.com>
Date: Sun, 17 Sep 2023 13:07:48 -0700
Subject: [PATCH] Fix llama.cpp double decoding

---
 modules/llamacpp_model.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/modules/llamacpp_model.py b/modules/llamacpp_model.py
index ed1f5b9e..44bb61f7 100644
--- a/modules/llamacpp_model.py
+++ b/modules/llamacpp_model.py
@@ -116,7 +116,7 @@ class LlamaCppModel:
         # Handle truncation
         prompt = self.encode(prompt)
         prompt = prompt[-get_max_prompt_length(state):]
-        prompt = self.decode(prompt).decode('utf-8')
+        prompt = self.decode(prompt)
 
         logit_processors = LogitsProcessorList()
         if state['ban_eos_token']: