https://github.com/ctongfei/nexus from @ctongfei looks like a more serious, comprehensive attempt to do something similar to what I was playing with in https://github.com/avibryant/nda - statically typed tensor shapes in Scala for deep learning.
-
Show this thread
The key feature they both share is using named singleton types to represent axes, and making sure that tensor ops are on *semantically* compatible shapes vs just possibly-coincidentally-matching numeric shapes. IMO this is a key usability advance for tensor libs.
9:47 AM - 19 Mar 2018
0 replies
0 retweets
10 likes
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.