Update README.md
Browse files
README.md
CHANGED
|
@@ -85,6 +85,45 @@ model-index:
|
|
| 85 |
- type: wer
|
| 86 |
value: 12.1
|
| 87 |
name: Test WER
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 88 |
---
|
| 89 |
# DeCRED-base
|
| 90 |
This is a **174M encoder-decoder Ebranchformer model** trained with an decoder-centric regularization technique on 6,000 hours of open-source normalised English data.
|
|
|
|
| 85 |
- type: wer
|
| 86 |
value: 12.1
|
| 87 |
name: Test WER
|
| 88 |
+
- task:
|
| 89 |
+
type: automatic-speech-recognition
|
| 90 |
+
name: Automatic Speech Recognition
|
| 91 |
+
dataset:
|
| 92 |
+
name: FLEURS
|
| 93 |
+
type: google/fleurs
|
| 94 |
+
split: test
|
| 95 |
+
args:
|
| 96 |
+
language: en_us
|
| 97 |
+
metrics:
|
| 98 |
+
- type: wer
|
| 99 |
+
value: 6.8
|
| 100 |
+
name: Test WER
|
| 101 |
+
- task:
|
| 102 |
+
type: automatic-speech-recognition
|
| 103 |
+
name: Automatic Speech Recognition
|
| 104 |
+
dataset:
|
| 105 |
+
name: Switchboard
|
| 106 |
+
type: unk
|
| 107 |
+
split: eval2000
|
| 108 |
+
args:
|
| 109 |
+
language: en
|
| 110 |
+
metrics:
|
| 111 |
+
- type: wer
|
| 112 |
+
value: 6.8
|
| 113 |
+
name: Test WER
|
| 114 |
+
- task:
|
| 115 |
+
type: automatic-speech-recognition
|
| 116 |
+
name: Automatic Speech Recognition
|
| 117 |
+
dataset:
|
| 118 |
+
name: Wall Street Journal
|
| 119 |
+
type: unk
|
| 120 |
+
split: eval92
|
| 121 |
+
args:
|
| 122 |
+
language: en
|
| 123 |
+
metrics:
|
| 124 |
+
- type: wer
|
| 125 |
+
value: 1.3
|
| 126 |
+
name: Test WER
|
| 127 |
---
|
| 128 |
# DeCRED-base
|
| 129 |
This is a **174M encoder-decoder Ebranchformer model** trained with an decoder-centric regularization technique on 6,000 hours of open-source normalised English data.
|