From eab7fcecb9f453d17ed3ab9f61e87a112aaa7d66 Mon Sep 17 00:00:00 2001
From: YunLiu <55491388+KumoLiu@users.noreply.github.com>
Date: Mon, 15 Apr 2024 11:56:02 +0800
Subject: [PATCH] minor fix

Signed-off-by: YunLiu <55491388+KumoLiu@users.noreply.github.com>
---
 acceleration/distributed_training/distributed_training.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/acceleration/distributed_training/distributed_training.md b/acceleration/distributed_training/distributed_training.md
index 79b9bddec0..870bd28cda 100644
--- a/acceleration/distributed_training/distributed_training.md
+++ b/acceleration/distributed_training/distributed_training.md
@@ -12,11 +12,11 @@ root_dir="/path/to/your/directory"  # Change this to your desired directory
 
 ## Multi-GPU Training
 
-Users can set your `NUM_GPUS_PER_NODE`, `NUM_NODES`, `INDEX_CURRENT_NODE`, as well as `DIR_OF_TESTDATA` for the directory of the test dataset.
+Users can set your `NUM_GPUS_PER_NODE`, `NUM_NODES`, `INDEX_CURRENT_NODE`, as well as `DIR_OF_DATA` for the directory of the test dataset.
 Then users can execute the following command to start multi-GPU model training:
 
 ```
-torchrun --nproc_per_node=NUM_GPUS_PER_NODE --nnodes=NUM_NODES brats_training_ddp.py -d DIR_OF_TESTDATA
+torchrun --nproc_per_node=NUM_GPUS_PER_NODE --nnodes=NUM_NODES brats_training_ddp.py -d DIR_OF_DATA
 ```
 
 ## Multi-Node Training