[
https://issues.apache.org/jira/browse/ARROW-3802?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16823066#comment-16823066
]
Wes McKinney commented on ARROW-3802:
-------------------------------------
TensorFlow uses the {{Eigen::half}} type to deal with float16. I presume that
NumPy has its own implementations of things, since float16 is not implemented
in hardware?
> [C++] Cast from integer to half float not implemented
> -----------------------------------------------------
>
> Key: ARROW-3802
> URL: https://issues.apache.org/jira/browse/ARROW-3802
> Project: Apache Arrow
> Issue Type: Bug
> Components: C++
> Reporter: Romain François
> Priority: Major
> Fix For: 0.14.0
>
>
> {code:java}
> library(reticulate)
> pa <- import("pyarrow")
> pa$array(c(1,2,3))$cast(pa$float16())
> #> Error in py_call_impl(callable, dots$args, dots$keywords):
> ArrowNotImplementedError: No cast implemented from double to halffloat
> #>
> #> Detailed traceback:
> #> File "pyarrow/array.pxi", line 277, in pyarrow.lib.Array.cast
> (/Users/travis/build/BryanCutler/arrow-dist/arrow/python/build/temp.macosx-10.6-intel-3.6/lib.cxx:30459)
> #> File "pyarrow/error.pxi", line 85, in pyarrow.lib.check_status
> (/Users/travis/build/BryanCutler/arrow-dist/arrow/python/build/temp.macosx-10.6-intel-3.6/lib.cxx:8570)
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
