Hugging Face's logo

nvidia
/
canary-1b

Automatic Speech Recognition
NeMo
PyTorch
automatic-speech-translation
speech
audio
Transformer
FastConformer
Conformer
NeMo
hf-asr-leaderboard
Eval Results
Model card Files Community
26
steveheh commited on
Commit
4009427
•
1 Parent(s): 569b0b1

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +118 -5
README.md CHANGED
@@ -79,7 +79,7 @@ model-index:
79
  args:
80
  language: en
81
  metrics:
82
- - name: Test WER
83
  type: wer
84
  value: 7.97
85
  - task:
@@ -93,7 +93,7 @@ model-index:
93
  args:
94
  language: de
95
  metrics:
96
- - name: Test WER
97
  type: wer
98
  value: 4.61
99
  - task:
@@ -107,7 +107,7 @@ model-index:
107
  args:
108
  language: es
109
  metrics:
110
- - name: Test WER
111
  type: wer
112
  value: 3.99
113
  - task:
@@ -121,7 +121,7 @@ model-index:
121
  args:
122
  language: fr
123
  metrics:
124
- - name: Test WER
125
  type: wer
126
  value: 6.53
127
  - task:
@@ -135,9 +135,122 @@ model-index:
135
  args:
136
  language: en-de
137
  metrics:
138
- - name: Test BLEU
139
  type: bleu
140
  value: 22.66
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
141
  metrics:
142
  - wer
143
  - bleu
 
79
  args:
80
  language: en
81
  metrics:
82
+ - name: Test WER (En)
83
  type: wer
84
  value: 7.97
85
  - task:
 
93
  args:
94
  language: de
95
  metrics:
96
+ - name: Test WER (De)
97
  type: wer
98
  value: 4.61
99
  - task:
 
107
  args:
108
  language: es
109
  metrics:
110
+ - name: Test WER (ES)
111
  type: wer
112
  value: 3.99
113
  - task:
 
121
  args:
122
  language: fr
123
  metrics:
124
+ - name: Test WER (Fr)
125
  type: wer
126
  value: 6.53
127
  - task:
 
135
  args:
136
  language: en-de
137
  metrics:
138
+ - name: Test BLEU (En->De)
139
  type: bleu
140
  value: 22.66
141
+ - task:
142
+ type: Automatic Speech Translation
143
+ name: automatic-speech-translation
144
+ dataset:
145
+ name: FLEURS
146
+ type: google/fleurs
147
+ config: en_us
148
+ split: test
149
+ args:
150
+ language: en-de
151
+ metrics:
152
+ - name: Test BLEU (En->Es)
153
+ type: bleu
154
+ value: 41.11
155
+ - task:
156
+ type: Automatic Speech Translation
157
+ name: automatic-speech-translation
158
+ dataset:
159
+ name: FLEURS
160
+ type: google/fleurs
161
+ config: en_us
162
+ split: test
163
+ args:
164
+ language: en-de
165
+ metrics:
166
+ - name: Test BLEU (En->Fr)
167
+ type: bleu
168
+ value: 40.76
169
+ - task:
170
+ type: Automatic Speech Translation
171
+ name: automatic-speech-translation
172
+ dataset:
173
+ name: FLEURS
174
+ type: google/fleurs
175
+ config: de_de
176
+ split: test
177
+ args:
178
+ language: de-en
179
+ metrics:
180
+ - name: Test BLEU (De->En)
181
+ type: bleu
182
+ value: 32.64
183
+ - task:
184
+ type: Automatic Speech Translation
185
+ name: automatic-speech-translation
186
+ dataset:
187
+ name: FLEURS
188
+ type: google/fleurs
189
+ config: es_419
190
+ split: test
191
+ args:
192
+ language: es-en
193
+ metrics:
194
+ - name: Test BLEU (Es->En)
195
+ type: bleu
196
+ value: 32.15
197
+ - task:
198
+ type: Automatic Speech Translation
199
+ name: automatic-speech-translation
200
+ dataset:
201
+ name: FLEURS
202
+ type: google/fleurs
203
+ config: fr_fr
204
+ split: test
205
+ args:
206
+ language: fr-en
207
+ metrics:
208
+ - name: Test BLEU (Fr->En)
209
+ type: bleu
210
+ value: 23.57
211
+ - task:
212
+ type: Automatic Speech Translation
213
+ name: automatic-speech-translation
214
+ dataset:
215
+ name: COVOST
216
+ type: covost2
217
+ config: de_de
218
+ split: test
219
+ args:
220
+ language: de-en
221
+ metrics:
222
+ - name: Test BLEU (De->En)
223
+ type: bleu
224
+ value: 37.67
225
+ - task:
226
+ type: Automatic Speech Translation
227
+ name: automatic-speech-translation
228
+ dataset:
229
+ name: COVOST
230
+ type: covost2
231
+ config: es_419
232
+ split: test
233
+ args:
234
+ language: es-en
235
+ metrics:
236
+ - name: Test BLEU (Es->En)
237
+ type: bleu
238
+ value: 40.7
239
+ - task:
240
+ type: Automatic Speech Translation
241
+ name: automatic-speech-translation
242
+ dataset:
243
+ name: COVOST
244
+ type: covost2
245
+ config: fr_fr
246
+ split: test
247
+ args:
248
+ language: fr-en
249
+ metrics:
250
+ - name: Test BLEU (Fr->En)
251
+ type: bleu
252
+ value: 40.42
253
+
254
  metrics:
255
  - wer
256
  - bleu