3 ms·
Not only that but it's also great for the English models as well. Many models are trained using the libre speech dataset with 960h of labeled audio. What if yo
by savant_penguin 5y ago
Not only that but it's also great for the English models as well. Many models are trained using the libre speech dataset with 960h of labeled audio.
What if you trained with 960000 hours instead??
- garmaine 5y agoLabeled data is easy to come by in popular language though. Just get an Audible subscription.