[model] support LiquidAI's LFM2.5-VL vision-language model (#9729)

src/llamafactory/extras/constants.py

@@ -1506,6 +1506,17 @@ register_model_group(
 )
 
 
+register_model_group(
+    models={
+        "LFM2.5-VL-1.6B": {
+            DownloadSource.DEFAULT: "LiquidAI/LFM2.5-VL-1.6B",
+        },
+    },
+    template="lfm2_vl",
+    multimodal=True,
+)
+
+
 register_model_group(
     models={
         "Llama-7B": {