Update README.md
Browse files
README.md
CHANGED
|
@@ -17,7 +17,7 @@ language: vi
|
|
| 17 |
## Efficient Conformer Encoder
|
| 18 |
Inspired from previous works done in Automatic Speech Recognition and Computer Vision, the Efficient Conformer encoder is composed of three encoder stages where each stage comprises a number of Conformer blocks using grouped attention. The encoded sequence is progressively downsampled and projected to wider feature dimensions, lowering the amount of computation while achieving better performance. Grouped multi-head attention reduce attention complexity by grouping neighbouring time elements along the feature dimension before applying scaled dot-product attention.
|
| 19 |
|
| 20 |
-
<img src="
|
| 21 |
|
| 22 |
## Installation
|
| 23 |
Clone GitHub repository and set up environment
|
|
@@ -62,7 +62,7 @@ bash test.sh
|
|
| 62 |
tensorboard --logdir callback_path
|
| 63 |
```
|
| 64 |
|
| 65 |
-
<img src="
|
| 66 |
|
| 67 |
## LibriSpeech Performance
|
| 68 |
|
|
|
|
| 17 |
## Efficient Conformer Encoder
|
| 18 |
Inspired from previous works done in Automatic Speech Recognition and Computer Vision, the Efficient Conformer encoder is composed of three encoder stages where each stage comprises a number of Conformer blocks using grouped attention. The encoded sequence is progressively downsampled and projected to wider feature dimensions, lowering the amount of computation while achieving better performance. Grouped multi-head attention reduce attention complexity by grouping neighbouring time elements along the feature dimension before applying scaled dot-product attention.
|
| 19 |
|
| 20 |
+
<img src="EfficientConformer.jpg" width="50%"/>
|
| 21 |
|
| 22 |
## Installation
|
| 23 |
Clone GitHub repository and set up environment
|
|
|
|
| 62 |
tensorboard --logdir callback_path
|
| 63 |
```
|
| 64 |
|
| 65 |
+
<img src="logs.jpg"/>
|
| 66 |
|
| 67 |
## LibriSpeech Performance
|
| 68 |
|