Corrected checkpoint parameter

The checkpoint parameter is never used and will be ignored. It should be set to `config_name`, so that the [EvaluationModule](https://github.com/huggingface/evaluate/blob/5aa3982a9a8c86e506860e381d428a64b0cce73b/src/evaluate/module.py#L149) reads it.

README.md

@@ -44,7 +44,7 @@ This metric takes as input lists of predicted sentences and reference sentences:
 ### Inputs
 - **predictions** (`list` of `str`s): List of generated sentences to score.
 - **references** (`list` of `str`s): List of references to compare to.
-- **checkpoint** (`str`): BLEURT checkpoint. Will default to `BLEURT-tiny` if not specified. Other models that can be chosen are: `"bleurt-tiny-128"`, `"bleurt-tiny-512"`, `"bleurt-base-128"`, `"bleurt-base-512"`, `"bleurt-large-128"`, `"bleurt-large-512"`, `"BLEURT-20-D3"`, `"BLEURT-20-D6"`, `"BLEURT-20-D12"` and `"BLEURT-20"`. 
+- **config_name** (`str`): BLEURT checkpoint. Will default to `BLEURT-tiny` if not specified. Other models that can be chosen are: `"bleurt-tiny-128"`, `"bleurt-tiny-512"`, `"bleurt-base-128"`, `"bleurt-base-512"`, `"bleurt-large-128"`, `"bleurt-large-512"`, `"BLEURT-20-D3"`, `"BLEURT-20-D6"`, `"BLEURT-20-D12"` and `"BLEURT-20"`. 
 
 ### Output Values
 - **scores** : a `list` of scores, one per prediction.