Update README.md

README.md

@@ -146,12 +146,15 @@ The processor exposes `process_texts`, `process_images`, and `score_multi_vector
 
 ### Prerequisites
 
-Note our model currently requires `flash-attn` to be installed. If you run into slow image encoding issue, please check whether `flash-attn` is installed properly.
+We strongly suggest `flash-attn` to be installed. If not, please change to `attention_impl="sdpa"`
+
+Currently we only support torch==2.8.0, for higher pytorch version, please build flash attention manually, otherwise performance throughput could be low.
 
 ```bash
-pip install torch torchvision transformers pillow requests
+pip install torch==2.8.0 torchvision==0.23.0 --index-url https://download.pytorch.org/whl/cu128
+pip install transformers pillow requests
 pip install flash-attn --no-build-isolation
-````
+```
 
 ### Inference Code
 
@@ -336,10 +339,11 @@ We gratefully acknowledge the support of **[Tomoro AI](https://tomoro.ai/)**, a
 If you use this model, please cite:
 
 ```bibtex
-@misc{huang2025tomoro_colqwen3_embed,
-  title        = {TomoroAI/tomoro-colqwen3-embed},
-  author       = {Xin Huang and Kye Min Tan and Albert Phelps},
-  year         = {2025},
-  url = {https://huggingface.co/TomoroAI/tomoro-colqwen3-embed-8b}
+@misc{huang2025beyond,
+  author = {Huang, Xin and Tan, Kye Min},
+  title = {Beyond Text: Unlocking True Multimodal, End-to-end RAG with Tomoro ColQwen3},
+  year = {2025},
+  url = {https://tomoro.ai/insights/beyond-text-unlocking-true-multimodal-end-to-end-rag-with-tomoro-colqwen3},
+  publisher = {Tomoro.ai}
 }
 ```