From d788fb49bf1ff2f41e651ded4e5b788f2185caad Mon Sep 17 00:00:00 2001
From: samfundev <samdinner@hotmail.com>
Date: Sat, 24 Jun 2023 15:51:46 -0400
Subject: [PATCH] Only concatenate after all batches are done

---
 llama_cpp/llama.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/llama_cpp/llama.py b/llama_cpp/llama.py
index a0b2030..d367601 100644
--- a/llama_cpp/llama.py
+++ b/llama_cpp/llama.py
@@ -405,6 +405,7 @@ class Llama:
         """
         assert self.ctx is not None
         n_ctx = self._n_ctx
+        scores = []
         for i in range(0, len(tokens), self.n_batch):
             batch = tokens[i : min(len(tokens), i + self.n_batch)]
             n_past = min(n_ctx - len(batch), len(self._input_ids))
@@ -430,9 +431,8 @@ class Llama:
             logits_view = llama_cpp.llama_get_logits(self.ctx)
             logits = [logits_view[i * cols : (i + 1) * cols] for i in range(rows)]
             self.eval_logits.extend(logits)
-            self._scores: npt.NDArray[np.single] = np.concatenate(
-                (self._scores, np.array(logits, dtype=np.single)), axis=0
-            )
+            scores.append(np.array(logits, dtype=np.single))
+        self._scores = np.concatenate(scores)
 
     def _sample(
         self,