Update README.md

README.md

@@ -5,14 +5,14 @@ license: mit
 
 Gemma 2B with recurrent local attention with context length of up to 10M. Our implemenation uses **<32GB** of memory!
 
-![Graphic of our implementation context](./images/graphic.png)
+![Graphic of our implementation context](./graphic.png)
 
 **Features:**
 
 - 10M sequence length on Gemma 2B.
-- Runs on less then 32GB of memory.
-- Native inference on Apple Silicon using MLX.
-- Highly performing retrieval - needle in hay stack.
+- Runs on less than 32GB of memory.
+- Native inference optimized for cuda. 
+- Recurrent local attention for O(N) memory. 
 
 ## Quick Start