From 26de815253f437a91e29a91eb0c430efacdafe70 Mon Sep 17 00:00:00 2001
From: jingyaogong <gongjy.cs@qq.com>
Date: Thu, 10 Apr 2025 22:36:34 +0800
Subject: [PATCH] update mask

---
 model/dataset.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/model/dataset.py b/model/dataset.py
index dae04a7..acbf716 100644
--- a/model/dataset.py
+++ b/model/dataset.py
@@ -20,8 +20,8 @@ class VLMDataset(Dataset):
         self.max_length = max_length
         self.preprocess = preprocess
         self.image_token = image_special_token
-        self.bos_id = tokenizer('<s>assistant\n', add_special_tokens=False).input_ids
-        self.eos_id = tokenizer('</s>\n', add_special_tokens=False).input_ids
+        self.bos_id = tokenizer('<s>assistant', add_special_tokens=False).input_ids
+        self.eos_id = tokenizer('</s>', add_special_tokens=False).input_ids
 
     def __len__(self):
         return len(self.samples)