Automatic Speech Recognition
NeMo
speech-recognition
speech-to-text
stt
transcription
speech
audio
asr
arabic
arabic-asr
arabic-speech
dialectal-arabic
msa
modern-standard-arabic
gulf-arabic
egyptian-arabic
levantine-arabic
maghrebi-arabic
multi-dialect
nemo-toolkit
nvidia
fastconformer
fastconformer-ctc
conformer
ctc
efficient
lightweight
compact
on-device
edge
cpu-inference
real-time
low-latency
low-resource
Eval Results (legacy)
Instructions to use lemuralabs/lemura-arabic-asr-lite with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use lemuralabs/lemura-arabic-asr-lite with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("lemuralabs/lemura-arabic-asr-lite") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Expand discovery metadata
Browse files
README.md
CHANGED
|
@@ -3,24 +3,129 @@ library_name: nemo
|
|
| 3 |
license: cc-by-4.0
|
| 4 |
language:
|
| 5 |
- ar
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
pipeline_tag: automatic-speech-recognition
|
|
|
|
|
|
|
| 7 |
tags:
|
| 8 |
- automatic-speech-recognition
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
- asr
|
| 10 |
- arabic
|
| 11 |
- arabic-asr
|
|
|
|
| 12 |
- dialectal-arabic
|
|
|
|
|
|
|
| 13 |
- gulf-arabic
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
- fastconformer
|
|
|
|
|
|
|
| 15 |
- ctc
|
| 16 |
- efficient
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
---
|
| 18 |
|
| 19 |
<div align="center">
|
| 20 |
|
| 21 |
<img src="logo.png" width="180"/>
|
| 22 |
|
| 23 |
-
# lemura-arabic-asr
|
| 24 |
|
| 25 |
   
|
| 26 |
|
|
|
|
| 3 |
license: cc-by-4.0
|
| 4 |
language:
|
| 5 |
- ar
|
| 6 |
+
- arz
|
| 7 |
+
- apc
|
| 8 |
+
- ary
|
| 9 |
+
- afb
|
| 10 |
pipeline_tag: automatic-speech-recognition
|
| 11 |
+
metrics:
|
| 12 |
+
- wer
|
| 13 |
tags:
|
| 14 |
- automatic-speech-recognition
|
| 15 |
+
- speech-recognition
|
| 16 |
+
- speech-to-text
|
| 17 |
+
- stt
|
| 18 |
+
- transcription
|
| 19 |
+
- speech
|
| 20 |
+
- audio
|
| 21 |
- asr
|
| 22 |
- arabic
|
| 23 |
- arabic-asr
|
| 24 |
+
- arabic-speech
|
| 25 |
- dialectal-arabic
|
| 26 |
+
- msa
|
| 27 |
+
- modern-standard-arabic
|
| 28 |
- gulf-arabic
|
| 29 |
+
- egyptian-arabic
|
| 30 |
+
- levantine-arabic
|
| 31 |
+
- maghrebi-arabic
|
| 32 |
+
- multi-dialect
|
| 33 |
+
- nemo
|
| 34 |
+
- nemo-toolkit
|
| 35 |
+
- nvidia
|
| 36 |
- fastconformer
|
| 37 |
+
- fastconformer-ctc
|
| 38 |
+
- conformer
|
| 39 |
- ctc
|
| 40 |
- efficient
|
| 41 |
+
- lightweight
|
| 42 |
+
- compact
|
| 43 |
+
- on-device
|
| 44 |
+
- edge
|
| 45 |
+
- cpu-inference
|
| 46 |
+
- real-time
|
| 47 |
+
- low-latency
|
| 48 |
+
- low-resource
|
| 49 |
+
model-index:
|
| 50 |
+
- name: lemura-arabic-asr-lite
|
| 51 |
+
results:
|
| 52 |
+
- task:
|
| 53 |
+
type: automatic-speech-recognition
|
| 54 |
+
name: Automatic Speech Recognition
|
| 55 |
+
dataset:
|
| 56 |
+
name: Open Universal Arabic ASR Leaderboard (average of 6 sets)
|
| 57 |
+
type: open-universal-arabic-asr
|
| 58 |
+
metrics:
|
| 59 |
+
- type: wer
|
| 60 |
+
value: 25.08
|
| 61 |
+
name: Average WER
|
| 62 |
+
- task:
|
| 63 |
+
type: automatic-speech-recognition
|
| 64 |
+
name: Automatic Speech Recognition
|
| 65 |
+
dataset:
|
| 66 |
+
name: MASC (clean)
|
| 67 |
+
type: masc-clean
|
| 68 |
+
metrics:
|
| 69 |
+
- type: wer
|
| 70 |
+
value: 7.27
|
| 71 |
+
name: WER
|
| 72 |
+
- task:
|
| 73 |
+
type: automatic-speech-recognition
|
| 74 |
+
name: Automatic Speech Recognition
|
| 75 |
+
dataset:
|
| 76 |
+
name: Common Voice 18 (Arabic)
|
| 77 |
+
type: common-voice-18-ar
|
| 78 |
+
metrics:
|
| 79 |
+
- type: wer
|
| 80 |
+
value: 9.74
|
| 81 |
+
name: WER
|
| 82 |
+
- task:
|
| 83 |
+
type: automatic-speech-recognition
|
| 84 |
+
name: Automatic Speech Recognition
|
| 85 |
+
dataset:
|
| 86 |
+
name: MGB-2
|
| 87 |
+
type: mgb-2
|
| 88 |
+
metrics:
|
| 89 |
+
- type: wer
|
| 90 |
+
value: 14.33
|
| 91 |
+
name: WER
|
| 92 |
+
- task:
|
| 93 |
+
type: automatic-speech-recognition
|
| 94 |
+
name: Automatic Speech Recognition
|
| 95 |
+
dataset:
|
| 96 |
+
name: MASC (noisy)
|
| 97 |
+
type: masc-noisy
|
| 98 |
+
metrics:
|
| 99 |
+
- type: wer
|
| 100 |
+
value: 23.65
|
| 101 |
+
name: WER
|
| 102 |
+
- task:
|
| 103 |
+
type: automatic-speech-recognition
|
| 104 |
+
name: Automatic Speech Recognition
|
| 105 |
+
dataset:
|
| 106 |
+
name: SADA
|
| 107 |
+
type: sada
|
| 108 |
+
metrics:
|
| 109 |
+
- type: wer
|
| 110 |
+
value: 37.28
|
| 111 |
+
name: WER
|
| 112 |
+
- task:
|
| 113 |
+
type: automatic-speech-recognition
|
| 114 |
+
name: Automatic Speech Recognition
|
| 115 |
+
dataset:
|
| 116 |
+
name: Casablanca
|
| 117 |
+
type: casablanca
|
| 118 |
+
metrics:
|
| 119 |
+
- type: wer
|
| 120 |
+
value: 58.24
|
| 121 |
+
name: WER
|
| 122 |
---
|
| 123 |
|
| 124 |
<div align="center">
|
| 125 |
|
| 126 |
<img src="logo.png" width="180"/>
|
| 127 |
|
| 128 |
+
# lemura-arabic-asr-lite
|
| 129 |
|
| 130 |
   
|
| 131 |
|