Update README.md

README.md

@@ -23,7 +23,7 @@ datasets:
 - **Model Developers:** Red Hat (Neural Magic)
 
 This model is a quantized version of [RedHatAI/Sparse-Llama-3.1-8B-tldr-2of4](https://huggingface.co/RedHatAI/Sparse-Llama-3.1-8B-tldr-2of4), which is fine-tuned on the [trl-lib/tldr](https://huggingface.co/datasets/trl-lib/tldr) dataset.
-This sparse-quantized model recovers 100% of the BERTScore (0.366) obtained by the dense model [RedHatAI/Llama-3.1-8B-tldr](https://huggingface.co/RedHatAI/Llama-3.1-8B-tldr). 
+This sparse-quantized model recovers 100% of the BERTScore (0.366) obtained by the dense model [RedHatAI/Llama-3.1-8B-tldr](https://huggingface.co/RedHatAI/Llama-3.1-8B-tldr) while providing up to 1.6x speedup. 
 
 
 ## Deployment