Update README.md

README.md

@@ -7,7 +7,9 @@ Only intended for conceptual validation, however the expert models do not seem t
 
 There are 8 completely different expert models based on Qwen-7B / CausalLM, six of which are specific domain models that have seen 50~100 billion tokens, including: a Toolformer/Agent expert model, a multilingual translation expert model, a mathematics expert model, a visual expert model, a coding and computer expert model, and an uncensored knowledge model — together forming the MoE model along with Qwen-Chat and Qwen-Base.
 
-The initialization of the gate is based on the hidden state of the few-shot prompt input from each expert model and undergoes simple alignment training.
+The initialization of the gate is based on the hidden state of the few-shot prompt input from each expert model and undergoes simple alignment training, on flan/ orca style.
+
+For multimodal input, please use visual.bin/py, should be the same as Qwen-VL.
 
 Prompt format: ChatML