Merge pull request #3 from calops/doc-improve-screenshot

Author: calops

README.md

@@ -14,9 +14,11 @@ to inject the actual language used within the screen, enabling proper highlighti
 > trivial ones. Properly detecting more complex stuff requires some lua code, which is what is done in hmts.nvim.
 
 Here's an example of the difference it can make:
-| Default behavior | With hmts.nvim |
-|------------------|-------------------------|
-|![image](https://github.com/calops/hmts.nvim/assets/4097716/cb71e7e2-f85c-4380-b2ff-127738230d63)|![image](https://github.com/calops/hmts.nvim/assets/4097716/220cd490-9bf2-44e8-962e-59f548861cf4)|
+
+Default behavior | With hmts.nvim
+---|---
+![before](https://github.com/calops/hmts.nvim/assets/4097716/44ef5636-292e-4932-bcc7-8c6554fca86e)|![after](https://github.com/calops/hmts.nvim/assets/4097716/16c6a094-8a01-4e6c-b09f-573bb074d8a0)
+
 
 ## Requirements