Update README.md

README.md

@@ -19,7 +19,7 @@ The model has 1.3 billion parameters.
 * The output will appear in the console window (after some time. You may want to open the RAM inspector to see what is going on.)
 
 ## Information
-This is the float32 version so it requires a lot of RAM (16GB) and VRAM (8GB). With less RAM it may take a long time to load. In the future we may add the float16 or uint8 quantized versions.
+This is the float32 version so it requires a lot of RAM (16GB) and VRAM (8GB). With less RAM it may take a long time to load. In the future we may add the float16 or uint8 quantized versions. This version also doesn't use caching of previous iterations so it is not the optimum speed but the implementation is simpler.
 
 ## Example Input
 ```