From d120d46b834a26546c582bdbf871deb8edecdbba Mon Sep 17 00:00:00 2001
From: Yuan Gong <yuangongfdu@gmail.com>
Date: Sat, 6 Jan 2024 20:20:51 -0500
Subject: [PATCH] add more pretrained model, and low-resource training scripts

---
 README.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/README.md b/README.md
index ee6490d..449c682 100644
--- a/README.md
+++ b/README.md
@@ -430,7 +430,7 @@ If you have a question about the code, please create an issue.
 
 ## Required Computational Resources
 
-For LTU/LTU-AS training, we use 4 X A6000 (4 X 48GB=196GB VRAM). The code can be run on 1 X A6000 (or similar GPUs). To run on smaller GPUs, turn on model parallelism, we were able to run it on 4 X A5000 (4 X 24GB = 96GB) (see [LTU script]() and [LTU-AS script]()).
+For LTU/LTU-AS training, we use 4 X A6000 (4 X 48GB=196GB VRAM). The code can be run on 1 X A6000 (or similar GPUs). To run on smaller GPUs, turn on model parallelism, we were able to run it on 4 X A5000 (4 X 24GB = 96GB) (see [LTU script](https://github.com/YuanGongND/ltu/blob/main/src/ltu/train_script/finetune_toy_low_resource.sh) and [LTU-AS script](https://github.com/YuanGongND/ltu/blob/main/src/ltu_as/train_script/finetune_toy_low_resource.sh)).
 
 For inference, the minimal would be 2 X TitanX (2 X 12GB = 24GB) for LTU and 4 X TitanX (4 X 12GB = 48GB) for LTU-AS (as Whisper takes some memory). However, you can run inference on CPUs.