Update README.md

README.md

@@ -17,7 +17,7 @@ tags:
 </p>
 
 
-Official pretrained weights for the **Celo2-base** learned update rule: This variant uses the learned update rule for all parameters without any optimization harness. For better performance, see [celo2](https://huggingface.co/amoudgl/celo2) that uses Newton-Schulz orthogonalization and AdamW for biases/embeddings.
+Official pretrained weights for **Celo2-base** learned update rule: This variant uses the learned update rule for all parameters without any optimization harness. For better performance, see [celo2](https://huggingface.co/amoudgl/celo2) that uses Newton-Schulz orthogonalization and AdamW for biases/embeddings.
 
 
 ## Quickstart