Update README.md

README.md

@@ -20,7 +20,7 @@ library_name: vllm
 
 This is an FP8 dynamically quantized (W8A8) version of `OpenGVLab/InternVL3_5-38B`optimized for high-performance inference with *vLLM*.
 
-The quantization process uses a specialized recipe that preserves the model's core visual understanding capabilities while reducing the memory footprint by nearly 50%.
+The quantization process uses a specialized recipe that preserves the model's core visual understanding capabilities while reducing the memory footprint by nearly 40%.
 
 ## Just Run It (vLLM serve)