add hunyuan_ocr

src/models/qwen3vl/generate.rs

@@ -33,6 +33,7 @@ pub struct Qwen3VLGenerateModel<'a> {
     eos_token_id1: u32,
     eos_token_id2: u32,
     generation_config: Qwen3VLGenerationConfig,
+    model_name: String,
 }
 
 impl<'a> Qwen3VLGenerateModel<'a> {
@@ -60,6 +61,7 @@ impl<'a> Qwen3VLGenerateModel<'a> {
             eos_token_id1: generation_config.eos_token_id[0] as u32,
             eos_token_id2: generation_config.eos_token_id[1] as u32,
             generation_config,
+            model_name: "qwen3vl".to_string(),
         })
     }
 }
@@ -120,7 +122,7 @@ impl<'a> GenerateModel for Qwen3VLGenerateModel<'a> {
         }
         let res = self.tokenizer.token_decode(generate)?;
         self.qwen3_vl.clear_kv_cache();
-        let response = build_completion_response(res, "qwen3vl");
+        let response = build_completion_response(res, &self.model_name);
         Ok(response)
     }
 
@@ -201,7 +203,7 @@ impl<'a> GenerateModel for Qwen3VLGenerateModel<'a> {
                     continue;
                 }
                 error_tokens.clear();
-                let chunk = build_completion_chunk_response(decoded_token, "qwen3vl", None, None);
+                let chunk = build_completion_chunk_response(decoded_token, &self.model_name, None, None);
                 yield Ok(chunk);
                 if next_token == self.eos_token_id1 || next_token == self.eos_token_id2 {
                     break;