Update README.md

README.md

@@ -18,6 +18,8 @@ It was pre-trained on pairwise posts, and contrastive learning guided them to le
 We randomly noise the hashtags to avoid trivial representation.
 Please refers to https://github.com/albertan017/HICL for more details.
 
+![Alt text](encoder-train.png)
+
 ### Model Description
 
 <!-- Provide a longer summary of what this model is. -->
@@ -60,11 +62,6 @@ with torch.no_grad():
 
 We do not inforce semantic similarity.
 
-### Recommendations
-
-<!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
-
-N.A.
 
 ## Training Details