Update preprocessor_config.json to use image_processor_type

This pull request addresses a warning message that appears when using the `nlpconnect/vit-gpt2-image-captioning` model. The warning suggests updating the `preprocessor_config.json` file to use `image_processor_type` instead of `feature_extractor_type`.

Changes made:
- Renamed `feature_extractor_type` to `image_processor_type` in `preprocessor_config.json`.

This change should help to align the model's configuration with the latest practices in the `transformers` library and remove the warning message for users of the model.

preprocessor_config.json

@@ -1,7 +1,7 @@
 {
   "do_normalize": true,
   "do_resize": true,
-  "feature_extractor_type": "ViTFeatureExtractor",
+  "image_processor_type": "ViTFeatureExtractor",
   "image_mean": [
     0.5,
     0.5,