oovword commited on
Commit
acce082
·
verified ·
1 Parent(s): 6a9fff8

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +24 -7
README.md CHANGED
@@ -15,6 +15,27 @@ inference: true
15
  library_name: transformers
16
  tags:
17
  - speech-translation
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
18
  ---
19
 
20
  # Model Card
@@ -77,13 +98,9 @@ For the `elevenlabs` dataset, the Ukrainian text was generated by ChatGPT and la
77
  Ukrainian peech and transcripts in the ML Spoken Words dataset are the authentic human data; the English text is machine-translated from Ukrainian by Azure Translator.
78
  **NOTE:** English translations were not human-verified or proofread due to time limitations and, as such, may contain mistakes and inaccuracies.
79
 
80
- Total (train): 10390 samples
81
- Total (dev): 2058 samples
82
- Total (test): 2828 samples
83
-
84
- Total duration (train): 10 hours 45 minutes 12 seconds
85
- Total duration (dev): 1 hour 36 minutes 7 seconds
86
- Total duration (test): 3 hours 1 minute 28 seconds
87
 
88
  ## Training
89
 
 
15
  library_name: transformers
16
  tags:
17
  - speech-translation
18
+ model-index:
19
+ - name: uk2en-speech-translation
20
+ results:
21
+ - task:
22
+ type: speech-translation
23
+ dataset:
24
+ name: Half-Synthetic Speech Dataset for Ukrainian-to-English Translation
25
+ type: oovword/speech-translation-uk-en
26
+ metrics:
27
+ - name: BLEU
28
+ type: bleu
29
+ value: 22.34
30
+ - task:
31
+ type: translation, speech-translation
32
+ dataset:
33
+ name: Half-Synthetic Speech Dataset for Ukrainian-to-English Translation
34
+ type: oovword/speech-translation-uk-en
35
+ metrics:
36
+ - name: ChrF++
37
+ type: chrf
38
+ value: 48.1
39
  ---
40
 
41
  # Model Card
 
98
  Ukrainian peech and transcripts in the ML Spoken Words dataset are the authentic human data; the English text is machine-translated from Ukrainian by Azure Translator.
99
  **NOTE:** English translations were not human-verified or proofread due to time limitations and, as such, may contain mistakes and inaccuracies.
100
 
101
+ - Total (train): 10390 samples (10 hours 45 minutes 12 seconds)
102
+ - Total (dev): 2058 samples (1 hour 36 minutes 7 seconds)
103
+ - Total (test): 2828 samples (3 hours 1 minute 28 seconds)
 
 
 
 
104
 
105
  ## Training
106