add qwen2.5vl model

src/models/mod.rs

@@ -0,0 +1,11 @@
+pub mod qwen2_5vl;
+use anyhow::Result;
+use candle_core::{DType, Device};
+use openai_dive::v1::resources::chat::ChatCompletionParameters;
+
+pub trait GenerateModel {
+    fn init(path: &str, device: Option<&Device>, dtype: Option<DType>) -> Result<Self>
+    where
+        Self: Sized;
+    fn generate(&mut self, mes: ChatCompletionParameters) -> Result<String>;
+}