system HF staff commited on
Commit
ee6d245
1 Parent(s): c12ea60

Update files from the datasets library (from 1.13.3)

Browse files

Release notes: https://github.com/huggingface/datasets/releases/tag/1.13.3

Files changed (2) hide show
  1. README.md +5 -0
  2. timit_asr.py +2 -0
README.md CHANGED
@@ -77,6 +77,9 @@ A typical data point comprises the path to the audio file, usually called `file`
77
  ```
78
  {
79
  'file': '/data/TRAIN/DR4/MMDM0/SI681.WAV',
 
 
 
80
  'text': 'Would such an act of refusal be useful?',
81
  'phonetic_detail': [{'start': '0', 'stop': '1960', 'utterance': 'h#'},
82
  {'start': '1960', 'stop': '2466', 'utterance': 'w'},
@@ -130,6 +133,8 @@ A typical data point comprises the path to the audio file, usually called `file`
130
 
131
  - file: A path to the downloaded audio file in .wav format.
132
 
 
 
133
  - text: The transcription of the audio file.
134
 
135
  - phonetic_detail: The phonemes that make up the sentence. The PHONCODE.DOC contains a table of all the phonemic and phonetic symbols used in TIMIT lexicon.
 
77
  ```
78
  {
79
  'file': '/data/TRAIN/DR4/MMDM0/SI681.WAV',
80
+ 'audio': {'path': '/data/TRAIN/DR4/MMDM0/SI681.WAV',
81
+ 'array': array([-0.00048828, -0.00018311, -0.00137329, ..., 0.00079346, 0.00091553, 0.00085449], dtype=float32),
82
+ 'sampling_rate': 16000},
83
  'text': 'Would such an act of refusal be useful?',
84
  'phonetic_detail': [{'start': '0', 'stop': '1960', 'utterance': 'h#'},
85
  {'start': '1960', 'stop': '2466', 'utterance': 'w'},
 
133
 
134
  - file: A path to the downloaded audio file in .wav format.
135
 
136
+ - audio: A dictionary containing the path to the downloaded audio file, the decoded audio array, and the sampling rate. Note that when accessing the audio column: `dataset[0]["audio"]` the audio file is automatically decoded and resampled to `dataset.features["audio"].sampling_rate`. Decoding and resampling of a large number of audio files might take a significant amount of time. Thus it is important to first query the sample index before the `"audio"` column, *i.e.* `dataset[0]["audio"]` should **always** be preferred over `dataset["audio"][0]`.
137
+
138
  - text: The transcription of the audio file.
139
 
140
  - phonetic_detail: The phonemes that make up the sentence. The PHONCODE.DOC contains a table of all the phonemic and phonetic symbols used in TIMIT lexicon.
timit_asr.py CHANGED
@@ -77,6 +77,7 @@ class TimitASR(datasets.GeneratorBasedBuilder):
77
  features=datasets.Features(
78
  {
79
  "file": datasets.Value("string"),
 
80
  "text": datasets.Value("string"),
81
  "phonetic_detail": datasets.Sequence(
82
  {
@@ -163,6 +164,7 @@ class TimitASR(datasets.GeneratorBasedBuilder):
163
 
164
  example = {
165
  "file": wav_path,
 
166
  "text": transcript,
167
  "phonetic_detail": phonemes,
168
  "word_detail": words,
 
77
  features=datasets.Features(
78
  {
79
  "file": datasets.Value("string"),
80
+ "audio": datasets.features.Audio(sampling_rate=16_000),
81
  "text": datasets.Value("string"),
82
  "phonetic_detail": datasets.Sequence(
83
  {
 
164
 
165
  example = {
166
  "file": wav_path,
167
+ "audio": wav_path,
168
  "text": transcript,
169
  "phonetic_detail": phonemes,
170
  "word_detail": words,