Add model card

This PR creates a model card, to ensure users can find the relevant information about the model, such as the paper, the github link,
and how to use it. It also sets the right pipeline_tag, ensuring the model can be found at https://huggingface.co/models?pipeline_tag=text-generation.

README.md

@@ -0,0 +1,8 @@
+---
+library_name: transformers
+pipeline_tag: text-generation
+---
+
+TransMLA: Equivalently Transforms Group Query Attention into Multi-head Latent Attention. 
+
+Find the relevant code at https://github.com/fxmeng/TransMLA.