Update README.md

README.md

@@ -93,8 +93,6 @@ Useful arguments to `datasets.Audio()` are:
 - `sampling_rate` and `mono=True` to control the sampling rate and number of channels.
 - `decode=False` to skip decoding the audio and just get the MP3 file paths and contents.
 
-The `load_dataset` function also accepts a `columns` parameter, which can be useful for example if you want to skip downloading the audio (see the example below).
-
 ## Running the benchmark
 
 The evaluation is implemented in our [`alt-eval` package](https://github.com/audioshake/alt-eval):
@@ -121,7 +119,7 @@ compute_metrics(dataset["text"], transcriptions, languages=dataset["language"])
 ```
 Alternatively, if you already have transcriptions, you might prefer to skip loading the `audio` column:
 ```python
-dataset = load_dataset("jamendolyrics/jam-alt", revision="v1.2.0", split="test", columns=["name", "text", "language", "license_type"])
+dataset = load_dataset("jamendolyrics/jam-alt", revision="v1.2.0", split="test").remove_columns("audio")
 ```
 
 ## Citation