chore: reflect maintenance mode

Author: pagezyhf

docs/source/index.mdx

@@ -17,7 +17,16 @@ limitations under the License.
 # 🤗 Optimum TPU
 
 <Tip>
-If you are here to start using HuggingFace products on TPU, then you are in the right place!
+**🚧 Optimum-TPU is now in maintenance mode.**
+
+We’ll continue to welcome community contributions for minor bug fixes, documentation improvements, and lightweight maintenance tasks.
+
+While this project is no longer under active development, you can continue exploring TPU solutions with:
+
+- [TPU inference](./tutorials/inference_on_tpu) for inference
+- [🤗 Accelerate](https://github.com/huggingface/accelerate) for training
+
+
 </Tip>
 
 Optimum TPU provides all the necessary machinery to leverage and optimize AI workloads running on [Google Cloud TPU devices](https://cloud.google.com/tpu/docs). Optimum-TPU is a HuggingFace solution to optimize HuggingFace products for the TPU platform. This allows users to use HuggingFace features and easy-to-use libraries on TPU with the best performance. We currently optimize transformers and TGI and integrate [HuggingFace hub](https://huggingface.co/models) so you can access HuggingFace's large library of models.