You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
This repository has been archived by the owner on Dec 31, 2020. It is now read-only.
Make it possible to define TF graphs from scratch, in addition to existing support for using externally-developed models.
Because the object model of TensorFlow graphs is extensive and ever-growing, we should hesitate to introduce a Flink-specific DSL. We should consider creating or incorporating a pure Scala TensorFlow library where such a DSL could be developed.
The text was updated successfully, but these errors were encountered:
Make it possible to define TF graphs from scratch, in addition to existing support for using externally-developed models.
Because the object model of TensorFlow graphs is extensive and ever-growing, we should hesitate to introduce a Flink-specific DSL. We should consider creating or incorporating a pure Scala TensorFlow library where such a DSL could be developed.
The text was updated successfully, but these errors were encountered: