From c673bd2fc99fd4e4d44ea913f6ea33dc696719fb Mon Sep 17 00:00:00 2001 From: randaller Date: Sun, 19 Mar 2023 15:05:44 +0300 Subject: [PATCH] Update hf-training-example.py --- hf-training-example.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/hf-training-example.py b/hf-training-example.py index 2afb9d4..977d1a1 100644 --- a/hf-training-example.py +++ b/hf-training-example.py @@ -17,7 +17,7 @@ tokenizer = llamahf.LLaMATokenizer.from_pretrained(MODEL) model = llamahf.LLaMAForCausalLM.from_pretrained(MODEL).cpu() if tokenizer.pad_token is None: - tokenizer.add_special_tokens({'pad_token': '[PAD]'}) + tokenizer.add_special_tokens({'pad_token': '<|endoftext|>'}) model.resize_token_embeddings(len(tokenizer))