From 596732a98199de543f5c660c612b84adc18969a4 Mon Sep 17 00:00:00 2001
From: oobabooga <112222186+oobabooga@users.noreply.github.com>
Date: Fri, 17 Feb 2023 12:35:30 -0300
Subject: [PATCH] The soft prompt length must be considered here too

---
 server.py | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/server.py b/server.py
index 145f2ae1..181a3c9d 100644
--- a/server.py
+++ b/server.py
@@ -505,11 +505,17 @@ def clean_chat_message(text):
     return text
 
 def generate_chat_prompt(text, tokens, name1, name2, context, chat_prompt_size, impersonate=False):
+    global soft_prompt, soft_prompt_tensor
+
     text = clean_chat_message(text)
     rows = [f"{context.strip()}\n"]
     i = len(history['internal'])-1
     count = 0
+
+    if soft_prompt:
+        chat_prompt_size -= soft_prompt_tensor.shape[1]
     max_length = min(get_max_prompt_length(tokens), chat_prompt_size)
+
     while i >= 0 and len(encode(''.join(rows), tokens)[0]) < max_length:
         rows.insert(1, f"{name2}: {history['internal'][i][1].strip()}\n")
         count += 1