Tensorflow 2.0: models migration and new design https://buff.ly/2zvL9oI Great overview of upcoming TF changes. Personally I agree with the author and I'm not a huge fan of eager mode either. Is it just because I'm used to thinking in terms of a graph? Is eager really superior?
-
I understand that it’s “easier” using an imperative style, but that IMO comes at a cost. My experience is that an imperative style leads throwing stuff at the wall (e.g. notebooks) + iterate/debug. Op graphs force you to think hard about what the computation should look like.
I think the right API is an API that matches the mental models of end users. People assembling deep networks from existing blocks think in terms of graphs of layers, so that should be the API. People writing forward pass code think in imperative terms, so that should be the API
-
-
Previously people writing TF V1 graph code for a custom layer would do so *while working from imperative mental models*, and that would lead to a lot frustration. I agree that if you are used to graphs of ops, then it's a fine paradigm, but in practice most people aren't
-
Ultimately the issue is not so much eager vs graph, it's making sure that what users expect (preexisting mental models) matches how the framework looks like and what it does
- Show replies
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.