Minor readme improvement

README.md

@@ -227,7 +227,7 @@ Each instance is an image, its representation in bytes, a pre-computed embedding
 }
 ```
 
-**Note**: The data is stored in Parquet for better performance. The dataset, which was later converted to Parquet, was generated from the original files using [this script](./scripts/wit.py). Additionally, 120 examples from the original files have incorrectly formatted one or more of the following fields: `original_height`, `original_width`, `mime_type` and `caption_attribution_description`. The fixed versions of these examples that were used in the generation script can be found in [this script](./scripts/corrected_examples.py)
+**Note**: The dataset is stored in Parquet for better performance. This dataset was generated from the original files using [this script](./scripts/wit.py). Additionally, 120 examples from the original files have incorrectly formatted one or more of the following fields: `original_height`, `original_width`, `mime_type` and `caption_attribution_description`. The fixed versions of these examples that were used in the generation script can be found [here](./scripts/corrected_examples.py)
 
 ### Data Fields