From 5075c16fccae6a2ce8fc894caaaf93f013b4a6fc Mon Sep 17 00:00:00 2001
From: Andrei Betlen <abetlen@gmail.com>
Date: Tue, 4 Apr 2023 13:08:21 -0400
Subject: [PATCH] Bugfix: n_batch should always be <= n_ctx

---
 llama_cpp/llama.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/llama_cpp/llama.py b/llama_cpp/llama.py
index f222dfd..4840caf 100644
--- a/llama_cpp/llama.py
+++ b/llama_cpp/llama.py
@@ -68,7 +68,7 @@ class Llama:
             maxlen=self.last_n_tokens_size,
         )
         self.tokens_consumed = 0
-        self.n_batch = n_batch
+        self.n_batch = min(n_ctx, n_batch)
 
         self.n_threads = n_threads or multiprocessing.cpu_count()