Update `base_model_relation` to `finetune`

Hello!

## Pull Request overview
* Update `base_model_relation` to `finetune`

## Details
The Hugging Face Hub sees this model as a `quantized` model instead of a `finetune`, because it has `onnx` files, so it says:


Instead of listing this as a finetune. This PR should fix that :)

- Tom Aarsen

README.md

@@ -2,6 +2,7 @@
 library_name: transformers
 license: apache-2.0
 base_model: answerdotai/ModernBERT-base
+base_model_relation: finetune
 tags:
 - generated_from_trainer
 metrics: