Add explanation for model-parallelism of LISA

OptimalScale · Apr 2, 2024 · 28fe928 · 28fe928
1 parent 49454a9
commit 28fe928
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/README.md b/README.md
@@ -113,7 +113,7 @@ cd data && ./download.sh alpaca && cd -
 ```
 
 ### Finetuning (LISA)
-[LISA](https://arxiv.org/abs/2403.17919) is a memory-efficient finetuning algorithm that allows tradeoff between memory and the number of randomly unfreezed layers.
+[LISA](https://arxiv.org/abs/2403.17919) is a memory-efficient finetuning algorithm that allows tradeoff between memory and the number of randomly unfreezed layers. This script currently is only tested in single gpus. Please stay tuned for our latest updates :smile:
 ```sh
 cd data && ./download.sh alpaca && cd -