default to llama 3.2 for ollama

examples/llm_ollama.rs

@@ -4,7 +4,7 @@ use langchain_rust::{language_models::llm::LLM, llm::ollama::client::Ollama};
 #[cfg(feature = "ollama")]
 #[tokio::main]
 async fn main() {
-    let ollama = Ollama::default().with_model("llama3");
+    let ollama = Ollama::default().with_model("llama3.2");
 
     let response = ollama.invoke("Hi").await.unwrap();
     println!("{}", response);

src/llm/ollama/client.rs

@@ -24,8 +24,8 @@ pub struct Ollama {
     pub(crate) options: Option<GenerationOptions>,
 }
 
-/// [llama3](https://ollama.com/library/llama3) is a 8B parameters, 4.7GB model.
-const DEFAULT_MODEL: &str = "llama3";
+/// [llama3.2](https://ollama.com/library/llama3.2) is a 3B parameters, 2.0GB model.
+const DEFAULT_MODEL: &str = "llama3.2";
 
 impl Ollama {
     pub fn new<S: Into<String>>(
@@ -152,15 +152,15 @@ mod tests {
     #[tokio::test]
     #[ignore]
     async fn test_generate() {
-        let ollama = Ollama::default().with_model("llama3");
+        let ollama = Ollama::default().with_model("llama3.2");
         let response = ollama.invoke("Hey Macarena, ay").await.unwrap();
         println!("{}", response);
     }
 
     #[tokio::test]
     #[ignore]
     async fn test_stream() {
-        let ollama = Ollama::default().with_model("llama3");
+        let ollama = Ollama::default().with_model("llama3.2");
 
         let message = Message::new_human_message("Why does water boil at 100 degrees?");
         let mut stream = ollama.stream(&vec![message]).await.unwrap();

src/llm/ollama/openai.rs

@@ -12,7 +12,7 @@ const OLLAMA_API_BASE: &str = "http://localhost:11434/v1";
 /// ## Example
 ///
 /// ```rs
-/// let ollama = OpenAI::new(OllamaConfig::default()).with_model("llama3");
+/// let ollama = OpenAI::new(OllamaConfig::default()).with_model("llama3.2");
 /// let response = ollama.invoke("Say hello!").await.unwrap();
 /// ```
 #[derive(Clone, Debug, Deserialize)]