From 557e881fcc48ec36b7eda4ab8d653e15209ca9d8 Mon Sep 17 00:00:00 2001
From: Hamid Shojanazeri <hamid.nazeri2010@gmail.com>
Date: Thu, 20 Jul 2023 04:05:17 +0000
Subject: [PATCH] aliginng the pad token with HF latest
---
inference/chat_completion.py | 8 +++-----
1 file changed, 3 insertions(+), 5 deletions(-)
diff --git a/inference/chat_completion.py b/inference/chat_completion.py
index 274f381f..bc5311d6 100644
--- a/inference/chat_completion.py
+++ b/inference/chat_completion.py
@@ -62,13 +62,11 @@ def main(
tokenizer = LlamaTokenizer.from_pretrained(model_name)
tokenizer.add_special_tokens(
{
- "eos_token": "</s>",
- "bos_token": "</s>",
- "unk_token": "</s>",
- "pad_token": "[PAD]",
+
+ "pad_token": "<PAD>",
}
)
-
+
chats = format_tokens(dialogs, tokenizer)
with torch.no_grad():
--
GitLab