Automatic Speech Recognition
NeMo
PyTorch
Bambara
speech
audio
CTC
QuartzNet
Bambara
NeMo
Eval Results (legacy)
Instructions to use RobotsMali/stt-bm-quartznet15x5-v0 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- NeMo
How to use RobotsMali/stt-bm-quartznet15x5-v0 with NeMo:
import nemo.collections.asr as nemo_asr asr_model = nemo_asr.models.ASRModel.from_pretrained("RobotsMali/stt-bm-quartznet15x5-v0") transcriptions = asr_model.transcribe(["file.wav"]) - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -128,7 +128,7 @@ The performance of Automatic Speech Recognition models is measured using Word Er
|
|
| 128 |
|
| 129 |
| Benchmark | Decoding | WER (%) ↓ | CER (%) ↓ |
|
| 130 |
|---------------|----------|-----------------|-----------------|
|
| 131 |
-
| Bam ASR Early | CTC |
|
| 132 |
| Nyana Eval | CTC | 65.42 | 30.66 |
|
| 133 |
|
| 134 |
These are **greedy WER numbers without external LM**.
|
|
|
|
| 128 |
|
| 129 |
| Benchmark | Decoding | WER (%) ↓ | CER (%) ↓ |
|
| 130 |
|---------------|----------|-----------------|-----------------|
|
| 131 |
+
| Bam ASR Early | CTC | 46.66 | 21.65 |
|
| 132 |
| Nyana Eval | CTC | 65.42 | 30.66 |
|
| 133 |
|
| 134 |
These are **greedy WER numbers without external LM**.
|