Remove Dual RoPE caches info and add litert-torch to README

Author: kamalkraj

litert_torch/generative/examples/embedding_gemma/README.md

@@ -14,7 +14,7 @@ EmbeddingGemma-300M is an encoder-only model with 24 layers. It uses a combinati
 To run this example and verify the results, you need the following packages:
 
 ```bash
-pip install transformers sentence-transformers safetensors
+pip install litert-torch transformers sentence-transformers safetensors
 ```
 
 ## Convert to TFLite
@@ -49,4 +49,4 @@ python verify.py
   --prompts="This is an example sentence."
 ```
 
-The verification script compares the final embeddings produced by the original `sentence-transformers` implementation and the reauthored `litert_torch` implementation to ensure parity.
+The verification script compares the final embeddings produced by the original `sentence-transformers` implementation and the reauthored `litert_torch` implementation to ensure parity.