Use image first in prompts (#8941)

* Use best practice in multimodal prompts to put the image first, the string prompt second

* Revert "Use best practice in multimodal prompts to put the image first, the string prompt second"

This reverts commit 594a259.

* Use best practice in multimodal prompts to put the image first, the string prompt second

Author: glaforge

vertexai/snippets/src/main/java/vertexai/gemini/SingleTurnMultimodal.java

@@ -61,8 +61,8 @@ public static void generateContent(String projectId, String location, String mod
       GenerativeModel model = new GenerativeModel(modelName, generationConfig, vertexAI);
       ResponseStream<GenerateContentResponse> responseStream = model.generateContentStream(
           ContentMaker.fromMultiModalData(
-              textPrompt,
-              PartMaker.fromMimeTypeAndData("image/jpg", decodedImage)
+              PartMaker.fromMimeTypeAndData("image/jpg", decodedImage),
+              textPrompt
           ));
       responseStream.stream().forEach(System.out::println);
     }