Spaces:

parler-tts

/

parler_tts

like 732

Running on Zero

Input text:

"What’s happening inside my application?" - an age-old question bothering anyone who deploys a software service. Packaging source code for an application makes it a black box for its users who can only interact with it through explicitly available APIs. Fortunately, we’ve had several developments in the field of observability in recent years that help us peek into this black box and react to anomalies.

Description:

Laura's voice is monotone yet slightly fast in delivery, with a very close recording that almost has no background noise.

With this input the large checkpoint is retuning an audio with:

• "What's" => "What better"
• "Software" => "forth"
• "Packaging source" => "Breadth is source"
• "APIs" => "cap keys".

It skips the last sentence completely in the live demo (not sure if due to a timeout).
The smaller model also has a problem with similar tokens along with "we've" in the last sentence.
It's hard for me to transcribe the errors for the smaller model into words.

Hope this helps in improving the model, especially the handling of what's and we've.