Automatic Speech Recognition
NeMo
PyTorch
4 languages
automatic-speech-translation
speech
audio
Transformer
FastConformer
Conformer
NeMo
hf-asr-leaderboard
Eval Results
nithinraok commited on
Commit
0624af3
1 Parent(s): b9993c0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +18 -0
README.md CHANGED
@@ -496,7 +496,25 @@ BLEU score on [mExpresso](https://huggingface.co/facebook/seamless-expressive#me
496
  |:-----------:|:---------:|:----------:|:----------:|:----------:|
497
  | 1.23.0 | canary-1b | 23.84 | 35.74 | 28.29 |
498
 
 
499
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
500
 
501
  ## NVIDIA Riva: Deployment
502
 
 
496
  |:-----------:|:---------:|:----------:|:----------:|:----------:|
497
  | 1.23.0 | canary-1b | 23.84 | 35.74 | 28.29 |
498
 
499
+ ## Model Fairness Evaluation
500
 
501
+ As outlined in the paper "Towards Measuring Fairness in AI: the Casual Conversations Dataset", we assessed the parakeet-tdt-1.1b model for fairness. The model was evaluated on the CausalConversations-v1 dataset, and the results are reported as follows:
502
+
503
+ ### Gender Bias:
504
+
505
+ | Gender | Male | Female | N/A | Other |
506
+ | :--- | :--- | :--- | :--- | :--- |
507
+ | Num utterances | 19325 | 24532 | 926 | 33 |
508
+ | % WER | 14.64 | 12.92 | 17.88 | 126.92 |
509
+
510
+ ### Age Bias:
511
+
512
+ | Age Group | (18-30) | (31-45) | (46-85) | (1-100) |
513
+ | :--- | :--- | :--- | :--- | :--- |
514
+ | Num utterances | 15956 | 14585 | 13349 | 43890 |
515
+ | % WER | 14.64 | 13.07 | 13.47 | 13.76 |
516
+
517
+ (Error rates for fairness evaluation are determined by normalizing both the reference and predicted text, similar to the methods used in the evaluations found at https://github.com/huggingface/open_asr_leaderboard.)
518
 
519
  ## NVIDIA Riva: Deployment
520