From 8ac44ef3bee1ed09f3dd4ad45b6f2769a4b9f0c8 Mon Sep 17 00:00:00 2001
From: Matthias Reso <13337103+mreso@users.noreply.github.com>
Date: Thu, 14 Sep 2023 14:26:23 +0000
Subject: [PATCH] Fix vocab size mismatch in inference due to added pad token
---
examples/inference.py | 8 +-------
1 file changed, 1 insertion(+), 7 deletions(-)
diff --git a/examples/inference.py b/examples/inference.py
index 28952920..ab4e7139 100644
--- a/examples/inference.py
+++ b/examples/inference.py
@@ -72,13 +72,7 @@ def main(
print("Module 'optimum' not found. Please install 'optimum' it before proceeding.")
tokenizer = LlamaTokenizer.from_pretrained(model_name)
- tokenizer.add_special_tokens(
- {
-
- "pad_token": "<PAD>",
- }
- )
- model.resize_token_embeddings(model.config.vocab_size + 1)
+ tokenizer.pad_token = tokenizer.eos_token
safety_checker = get_safety_checker(enable_azure_content_safety,
enable_sensitive_topics,
--
GitLab