Hugging Face's logo

Datasets:
davidggphy
/
librispeech-arpabet-processed

Tasks:
Automatic Speech Recognition
Modalities:
Audio
Text
Time-series
Formats:
parquet
Languages:
English
Size:
10K - 100K
Tags:
phoneme-recognition
arpabet
pronunciation
wav2vec2
ctc
speech
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
Community
1
davidggphy commited on
Commit
e65afe3
·
verified ·
1 Parent(s): 896c382

Upload README.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. README.md +54 -23
README.md CHANGED
@@ -14,23 +14,55 @@ tags:
14
  pretty_name: LibriSpeech ARPAbet Phonemes
15
  size_categories:
16
  - 10K<n<100K
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
  dataset_info:
18
- features:
19
- - name: audio
20
- dtype: audio
21
- - name: input_values
22
- sequence: float32
23
- - name: labels
24
- sequence: int32
25
- - name: text
26
- dtype: string
27
- - name: duration
28
- dtype: float32
29
- splits:
30
- - name: train
31
- num_examples: 15928
32
- - name: test
33
- num_examples: 1770
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
34
  ---
35
 
36
  # LibriSpeech ARPAbet Processed Dataset
@@ -80,15 +112,14 @@ from datasets import load_dataset
80
  from huggingface_hub import hf_hub_download
81
  import json
82
 
83
- # Load the full dataset
84
- dataset = load_dataset("davidggphy/librispeech-arpabet-processed")
85
 
86
- # Or load a subset (faster for testing)
87
- dataset = load_dataset("davidggphy/librispeech-arpabet-processed", split="train[:1000]") # First 1000
88
- dataset = load_dataset("davidggphy/librispeech-arpabet-processed", split="train[:10%]") # First 10%
89
 
90
- # Streaming mode (no full download)
91
- stream = load_dataset("davidggphy/librispeech-arpabet-processed", split="train", streaming=True)
92
  for sample in stream.take(100):
93
  print(sample["text"])
94
 
 
14
  pretty_name: LibriSpeech ARPAbet Phonemes
15
  size_categories:
16
  - 10K<n<100K
17
+ configs:
18
+ - config_name: full
19
+ default: true
20
+ data_files:
21
+ - split: train
22
+ path: data/train/*.parquet
23
+ - split: test
24
+ path: data/test/*.parquet
25
+ - config_name: mini
26
+ data_files:
27
+ - split: train
28
+ path: data/mini/train.parquet
29
+ - split: test
30
+ path: data/mini/test.parquet
31
  dataset_info:
32
+ - config_name: full
33
+ features:
34
+ - name: audio
35
+ dtype: audio
36
+ - name: input_values
37
+ sequence: float32
38
+ - name: labels
39
+ sequence: int32
40
+ - name: text
41
+ dtype: string
42
+ - name: duration
43
+ dtype: float32
44
+ splits:
45
+ - name: train
46
+ num_examples: 15928
47
+ - name: test
48
+ num_examples: 1770
49
+ - config_name: mini
50
+ features:
51
+ - name: audio
52
+ dtype: audio
53
+ - name: input_values
54
+ sequence: float32
55
+ - name: labels
56
+ sequence: int32
57
+ - name: text
58
+ dtype: string
59
+ - name: duration
60
+ dtype: float32
61
+ splits:
62
+ - name: train
63
+ num_examples: 900
64
+ - name: test
65
+ num_examples: 100
66
  ---
67
 
68
  # LibriSpeech ARPAbet Processed Dataset
 
112
  from huggingface_hub import hf_hub_download
113
  import json
114
 
115
+ # Load full dataset (~17k samples, ~15GB)
116
+ dataset = load_dataset("davidggphy/librispeech-arpabet-processed", "full")
117
 
118
+ # Load mini dataset (1000 samples, ~1GB) - great for quick testing!
119
+ dataset = load_dataset("davidggphy/librispeech-arpabet-processed", "mini")
 
120
 
121
+ # Streaming mode (download samples on-demand)
122
+ stream = load_dataset("davidggphy/librispeech-arpabet-processed", "full", split="train", streaming=True)
123
  for sample in stream.take(100):
124
  print(sample["text"])
125