NunoSantos commited on
Commit
380417a
1 Parent(s): ce7433e

Upload config

Browse files
Files changed (1) hide show
  1. readme_template.md +113 -0
readme_template.md ADDED
@@ -0,0 +1,113 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ language:
3
+ - en
4
+ license: cc-by-4.0
5
+ library_name: nemo
6
+ datasets:
7
+ - librispeech_asr
8
+ - mozilla-foundation/common_voice_7_0
9
+ - vctk
10
+ - fisher_corpus
11
+ - Switchboard-1
12
+ - WSJ-0
13
+ - WSJ-1
14
+ - National-Singapore-Corpus-Part-1
15
+ - National-Singapore-Corpus-Part-6
16
+ - VoxPopuli-(EN)
17
+ - Europarl-ASR-(EN)
18
+ - Multilingual-LibriSpeech-(2000-hours)
19
+ thumbnail: null
20
+ tags:
21
+ - automatic-speech-recognition
22
+ - speech
23
+ - audio
24
+ - CTC
25
+ - Conformer
26
+ - Transformer
27
+ - NeMo
28
+ - pytorch
29
+ model-index:
30
+ - name: stt_en_conformer_ctc_small
31
+ results: []
32
+
33
+ ---
34
+
35
+
36
+ ## Model Overview
37
+
38
+ <DESCRIBE IN ONE LINE THE MODEL AND ITS USE>
39
+
40
+ ## NVIDIA NeMo: Training
41
+
42
+ To train, fine-tune or play with the model you will need to install [NVIDIA NeMo](https://github.com/NVIDIA/NeMo). We recommend you install it after you've installed latest Pytorch version.
43
+ ```
44
+ pip install nemo_toolkit['all']
45
+ ```
46
+
47
+ ## How to Use this Model
48
+
49
+ The model is available for use in the NeMo toolkit [3], and can be used as a pre-trained checkpoint for inference or for fine-tuning on another dataset.
50
+
51
+ ### Automatically instantiate the model
52
+
53
+ ```python
54
+ import nemo.collections.asr as nemo_asr
55
+ asr_model = nemo_asr.models.ASRModel.from_pretrained("NunoSantos/stt_en_conformer_ctc_small")
56
+ ```
57
+
58
+ ### Transcribing using Python
59
+ First, let's get a sample
60
+ ```
61
+ wget https://dldata-public.s3.us-east-2.amazonaws.com/2086-149220-0033.wav
62
+ ```
63
+ Then simply do:
64
+ ```
65
+ asr_model.transcribe(['2086-149220-0033.wav'])
66
+ ```
67
+
68
+ ### Transcribing many audio files
69
+
70
+ ```shell
71
+ python [NEMO_GIT_FOLDER]/examples/asr/transcribe_speech.py pretrained_name="NunoSantos/stt_en_conformer_ctc_small" audio_dir="<DIRECTORY CONTAINING AUDIO FILES>"
72
+ ```
73
+
74
+ ### Input
75
+
76
+ This model accepts 16000 KHz Mono-channel Audio (wav files) as input.
77
+
78
+ ### Output
79
+
80
+ This model provides transcribed speech as a string for a given audio sample.
81
+
82
+ ## Model Architecture
83
+
84
+ <ADD SOME INFORMATION ABOUT THE ARCHITECTURE>
85
+
86
+ ## Training
87
+
88
+ <ADD INFORMATION ABOUT HOW THE MODEL WAS TRAINED - HOW MANY EPOCHS, AMOUNT OF COMPUTE ETC>
89
+
90
+ ### Datasets
91
+
92
+ <LIST THE NAME AND SPLITS OF DATASETS USED TO TRAIN THIS MODEL (ALONG WITH LANGUAGE AND ANY ADDITIONAL INFORMATION)>
93
+
94
+ ## Performance
95
+
96
+ <LIST THE SCORES OF THE MODEL -
97
+ OR
98
+ USE THE Hugging Face Evaluate LiBRARY TO UPLOAD METRICS>
99
+
100
+ ## Limitations
101
+
102
+ <DECLARE ANY POTENTIAL LIMITATIONS OF THE MODEL>
103
+
104
+ Eg:
105
+ Since this model was trained on publicly available speech datasets, the performance of this model might degrade for speech which includes technical terms, or vernacular that the model has not been trained on. The model might also perform worse for accented speech.
106
+
107
+
108
+ ## References
109
+
110
+ <ADD ANY REFERENCES HERE AS NEEDED>
111
+
112
+ [1] [NVIDIA NeMo Toolkit](https://github.com/NVIDIA/NeMo)
113
+