From 920c0e82177ceab3a0e574f68b1184ad9a16995c Mon Sep 17 00:00:00 2001
From: Kexin Wang <71278644+kwang2049@users.noreply.github.com>
Date: Fri, 17 Dec 2021 16:40:22 +0100
Subject: [PATCH] Update README.md

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

diff --git a/gpl/toolkit/README.md b/gpl/toolkit/README.md
index 51abb2f..d53c6f6 100644
--- a/gpl/toolkit/README.md
+++ b/gpl/toolkit/README.md
@@ -9,7 +9,7 @@ This is because:
 However, the last layer of "facebook/dpr-question_encoder-single-nq-base" is actually a dense linear (with an activation function).
 
 To start from these checkpoints, one need to transform them into SBERT-format. We here provide two examples in [reformat.py](./reformat.py) to transform `simcse_like` and `dpr_like` models into SBERT-format. For example, one can run:
-```python
+```bash
 python -m gpl.toolkit.reformat \
     --template "dpr_like" \
     --model_name_or_path "facebook/dpr-question_encoder-single-nq-base" \