From 84d8613f1ff8db40da2ae7a9bc476dca033752e0 Mon Sep 17 00:00:00 2001 From: Quentin Anthony Date: Sun, 22 Sep 2024 14:37:12 -0700 Subject: [PATCH] fix typo from https://github.com/EleutherAI/gpt-neox/pull/1244/files#diff-383134de6f3512484e20625419bd5fb6b1675a922f47aeb1a6bd3cff6185a754R126 --- megatron/data/gpt2_dataset.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/megatron/data/gpt2_dataset.py b/megatron/data/gpt2_dataset.py index e37c558d2..c4729cc3e 100644 --- a/megatron/data/gpt2_dataset.py +++ b/megatron/data/gpt2_dataset.py @@ -123,7 +123,7 @@ def __getitem__(self, idx): samples.append( dataset.get( self.doc_idx[doc_index_f], - offset=offset_l, + offset=offset_f, length=offset_l - offset_f + 1, ) )