Update processor_config.json

Adds `num_additional_image_tokens` so that the model can run with the latest transformers version. Same modification as in https://huggingface.co/unsloth/llava-1.5-7b-hf-bnb-4bit/discussions/1.
Some context for this change can be read in https://github.com/huggingface/transformers/pull/35779

processor_config.json

@@ -2,5 +2,6 @@
   "image_token": "<image>",
   "patch_size": 14,
   "processor_class": "LlavaNextProcessor",
-  "vision_feature_select_strategy": "default"
+  "vision_feature_select_strategy": "default",
+  "num_additional_image_tokens": 1
 }