Closed watercrossing closed 9 years ago
Yep, need to create our own copy around it for now. Note this linear chain representation is deprecated in that now we have variable length sequence domains that allow us to use the same domain across several instances. This allows to use same term (and memory allocation) for all instances in the corpus, which makes things more memory and cpu efficient
I think this outdated now, but I will create a related issue
https://github.com/wolfe-pack/wolfe/blob/dev/wolfe-examples/src/main/scala/ml/wolfe/examples/LinearChain.scala fails with the error message:
It seems like we are using a SingletonTensor in an unsupported way?