Bill Peebles

Hi @AK391, thanks for the suggestion! It would be cool to host a demo on Huggingface. Do you know if it's possible to deploy demos to Spaces with a GPU...

You can use our portable implementations of the Hessian Penalty for fine-tuning StyleGAN(-v2) in either PyTorch or TensorFlow. While it should usually be fine to directly apply the Hessian Penalty...

Thanks for the question. I have observed that it sometimes decreases quickly when training, but not to 1e-8. Is your MSE decreasing slowly? It may be that the hessian penalty...

It does seem slightly odd to me that it drops so low so quickly. Does your plot include the value of the hessian penalty at initialization? Assuming it doesn't start...

Also I would recommend against using `tf.abs` in the reduction function (sorry for any confusion about this, it was only intended to be used as a test case). This may...

Sorry for the delay in responding. I think the behavior you're describing is more typical of what I've seen applying it to ProGAN. 0.01 magnitude can be reasonable depending on...

Thanks for the amazing repo! Do you guys by chance have any update on when multi-agent PettingZoo atari will be integrated?

I'm experiencing the same issue running the MPII example with PyTorch 0.3.1, a 1080 and with default parameters except stacks=1.