Hacker News Reader

Is this device optimized for forward passes or backward passes or both?

It seems to me that Google engineers could use Tesla's or other high end GPU's for training and development, but then deploy those models on hardware optimized for forward passes...

View on HN

If "forward passes" means inference (as opposed to training), then the post says the first generation targets that. I don't think they say anything about any future generations (other than they're working on them).

I mean, there's no real reason they shouldn't be able to do a backwards pass assuming they're using trivially differentiable activation functions.

Reply on news.ycombinator.com