chapters/en/chapter5/asr_models.mdx has wrong code.

https://github.com/huggingface/audio-transformers-course/blob/4bc95b750205face0900f796f04bae58c54384b5/chapters/en/chapter5/asr_models.mdx?plain=1#L213-L216

When this code runs.
```
ValueError: Bad split: validation. Available splits: ['dev', 'test', 'train', '9_hours', '1_hours']
```
and the split param should be changed to use streaming mode.

and after changed `split="validation"` to something else (ex: `split="dev"`)

and the next code . it's not work. but `sample["transcript"]` works.
```
print(sample["text"])
```
the print(sample) shown like below
```
{'audio': {'path': '10367_10282_000000.opus', 'array': array([0.00126664, 0.00119767, 0.00108108, ..., 0.00066889, 0.00055418,
0.00020247]), 'sampling_rate': 16000}, 'original_path': 'http://www.archive.org/download/isaias_1603_librivox/isaias_28_reinavalera_64kb.mp3', 'begin_time': 193.17, 'end_time': 205.18, 'transcript': 'y á los hijos de los extranjeros que se allegaren á jehová para ministrarle y que amaren el nombre de jehová para ser sus siervos á todos los que guardaren el sábado de profanarlo y abrazaren mi pacto', 'audio_duration': 12.01000000000002, 'speaker_id': '10367', 'chapter_id': '10282', 'file': '10367_10282_000000.opus', 'id': '10367_10282_000000'}
```

The datasets might changed after wrote this chapter page.

Open chapters/en/chapter5/asr_models.mdx around lines 213-216 and reproduce the dataset-loading example with the current datasets available. Check the valid split names and streaming behavior, then verify the following sample access uses the field present in the returned record. Done when the chapter example runs without the split error and prints the sample transcript.

Written by the indexing model from the issue text.