huling0

5 ms·

1.
by huling0 7y ago
though I have to admit this looks like the best in class, there is a public real time deep fake service already available: https://www.aphrodite.ai/
2.
by huling0 8y ago
http://notarealhuman.com/ has cats as well :)
3.
by huling0 8y ago
good suggestion. Looking into latent vector manipulation now.
4.
by huling0 8y ago
Why this negativity? It's not harming anyone. Just fun to play with the new StyleGAN network which is only out since a few weeks.
5.
by huling0 8y ago
Thanks! I missed that one - seems to do the same :-)
6.
by huling0 8y ago
It is based on https://arxiv.org/pdf/1812.04948.pdf Images can easily be generated on the fly. Inference time is ~100ms.
7.
Show HN: Fake Profile Picture Generator(notarealhuman.com)
6 points by huling0 8y ago|9 comments
8.
by huling0 8y ago
Care to share what weekly subscription newspaper you read?
9.
by huling0 8y ago
I recently ran into that issue - I have a nice fine-tuned yocto image and now a customer wants to "log in via ssh and install a custom raspbian package". I was thinking of the docker solution as well.
10.
by huling0 8y ago
yeah, but does not look like something many hobbyists would try for fun ;-)
11.
by huling0 8y ago
Yes, but it just cuts out the face and pastes it on a different person/background. It does not do full reenactment where you keep the entire target video environment.
12.
by huling0 8y ago
From the paper: "we have to use all the GPUs in DGX1 (8 V100 GPUs, each with 16GB memory) for training. We distribute the generator computation task to 4 GPUs and the discriminator computation task to the other 4 GPUs. Training takes ∼
13.
by huling0 8y ago
One of the example translates a full human pose to a video of a dancer. If the network would be trained on the facial pose(?) / features only, would that recreate something like the facial reenactment in http://niessnerlab.o