From 9f991040388d7d0b3b36f24de64cb692503a8d71 Mon Sep 17 00:00:00 2001 From: Aman Karmani Date: Mon, 7 Aug 2023 01:04:56 -0700 Subject: [PATCH] update comment for group_by_length --- README.md | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/README.md b/README.md index fe22bbc31..479b1ab4a 100644 --- a/README.md +++ b/README.md @@ -426,7 +426,8 @@ save_safetensors: # whether to mask out or include the human's prompt from the training labels train_on_inputs: false -# don't use this, leads to wonky training (according to someone on the internet) +# group similarly sized data to minimize padding +# may be slower to start as it must download and sort the entire dataset group_by_length: false # Whether to use gradient checkpointing https://huggingface.co/docs/transformers/v4.18.0/en/performance#gradient-checkpointing