If I'm not mistaken, it seems that the current limitation is that it needs to be produced sequentially for a dependent sequence of audio, perhaps some independent sentences can be run simultaneously using copies of the net assuming no memory limitations. I wonder if it's already possible to create an auidobook for instance in reasonable time.
HN user
mdsteph
Co-founder at Syft Analytics
That's very cool, it's amazing what's become possible.
The morphing is impressive. I wonder if something similar can morph between faces. It may even be possible to morph faces of parents and get something similar to the children of the parents, especially if that's set as the objective. Looking forward to release of the code.
It may matter, thanks for the details, this will allow me to test and improve the frame rate on a similar device.
Thanks :) will definitely look at improving the speed, I think it might be a slight amount of image resizing which I can look to optimise away. Otherwise, just to baseline the performance, could I perhaps ask what device you tried it on?
I agree that non valid patents should be killed. If a patent is valid perhaps its usefulness in monetary terms to the community is still worth more than the present value of what trolls and legitimate patent holders could expect to earn by holding it.
I don't agree that killing patents is the right approach, how about this as an alternative: The site's admin team could identify patents of particularly useful or disruptive potential and users could pledge a value to bring this patent into the public domain, and their pledge could remain valid for some reasonable period of time, the site could then put forward the aggregate pledge and buy the patent with the view to releasing it. Perhaps a price could be agreed upon up front. The same could be done to bring creative works like books into the public domain.