From 8404fbeef1e5af4aef67a85da2189e6102803efe Mon Sep 17 00:00:00 2001
From: CharlesCNorton <135471798+CharlesCNorton@users.noreply.github.com>
Date: Fri, 10 Jan 2025 19:41:41 -0500
Subject: [PATCH] Update post_training readme.md

This commit fixes the redundancy in the sentence:

"This is accomplished by using utilizing Fully Sharded Data Parallel (FSDP) and Tensor Parallelism." to simply read "by utilizing Fully Sharded Data Parallel (FSDP) and Tensor Parallelism."
---
 cosmos1/models/diffusion/nemo/post_training/README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/cosmos1/models/diffusion/nemo/post_training/README.md b/cosmos1/models/diffusion/nemo/post_training/README.md
index ba8c3ded..a0d351a3 100644
--- a/cosmos1/models/diffusion/nemo/post_training/README.md
+++ b/cosmos1/models/diffusion/nemo/post_training/README.md
@@ -131,7 +131,7 @@ Executing the [data preprocessing script](./prepare_dataset.py) generates the fo
 
 ### 3. Post-train the Model
 
-The third step is to post-train the model. This step uses NeMo Framework's data and model parallelism capabilities to train the model on the post-training samples. This is accomplished by using utilizing Fully Sharded Data Parallel (FSDP) and Tensor Parallelism.
+The third step is to post-train the model. This step uses NeMo Framework's data and model parallelism capabilities to train the model on the post-training samples. This is accomplished by utilizing Fully Sharded Data Parallel (FSDP) and Tensor Parallelism.
 
 - **FSDP**: Distributes model parameters, optimizer states, and activations across all GPUs
 - **Tensor Parallelism**: Spreads the parameter tensor of individual layers across GPUs.