From 1594c1dade6845377554bb72b168f69185368d9e Mon Sep 17 00:00:00 2001
From: sayakpaul <spsayakpaul@gmail.com>
Date: Thu, 26 Dec 2024 08:12:41 +0530
Subject: [PATCH] deepspeed note

---
 README.md | 1 +
 1 file changed, 1 insertion(+)

diff --git a/README.md b/README.md
index 60dfc7f..ab56653 100644
--- a/README.md
+++ b/README.md
@@ -385,6 +385,7 @@ If you would like to use a custom dataset, refer to the dataset preparation guid
 
 > [!NOTE]
 > To lower memory requirements:
+> - Use a DeepSpeed config to launch training (refer to [`accelerate_configs/deepspeed.yaml`](./accelerate_configs/deepspeed.yaml) as an example).
 > - Pass `--precompute_conditions` when launching training.
 > - Pass `--gradient_checkpointing` when launching training.
 > - Do not perform validation/testing. This saves a significant amount of memory, which can be used to focus solely on training if you're on smaller VRAM GPUs.