Skip to content
 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Benchmarking - Evaluation Metrics

Time to First Audio (TTFA)

Measures the time (in milliseconds) from when a text input is sent to when the first segment of audio is played.

A lower TTFA is critical for interactive applications where quick responses are necessary.

See the attached code for exact details.

Word Error Rate (WER)

Evaluates the accuracy of the generated speech by comparing it against expected transcriptions.

Calculated as the percentage of words that are incorrect due to substitutions, omissions, or insertions.

Lower WER values indicate better speech accuracy.

Alphanumeric Performance

Measures the ability of the TTS model to correctly pronounce and articulate alphanumeric sequences, such as phone numbers, addresses, and serial codes.

Critical for voice assistants in customer support and transactional applications.

About

Voice AI benchmarks for all of the latest TTS and STT models.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages