The fact that evaluating ∇f(x) is as fast as f(x) is very important and often misunderstood http://timvieira.github.io/blog/post/2016/09/25/evaluating-fx-is-as-fast-as-fx/ …https://twitter.com/gabrielpeyre/status/1167663307668373504 …
-
-
Beyond what I wrote up in the post?
-
got it. I thought misunderstanding was on computing the gradient not on the asymptotic complexity.
- Još 1 odgovor
Novi razgovor -
-
-
@xtimv post is great and highlights the key aspect of mitigating time vs memory, which is the most crucial aspect. Advanced methods to automatically adress this seem to be lacking in@PyTorch and@TensorFlow though. -
Thanks! It means a lot coming from you - I'm a huge fan of your account!
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.