Oracle offers GraphPipe spec for machine learning data transmission

(Foto: Istock)

Oracle offers GraphPipe spec for machine learning data transmission

IDG NEWS: GraphPipe is intended to bring the efficiency of a binary, memory-mapped format while being simple and light on dependencies

Vil du fortsette å lese, velg et av alternativene nedenfor

  • Logg inn!

    Du har abonnement og er registrert som bruker.

  • Har abonnement!

    Du har abonnement, men ikke registrert deg.

  • Bestill abonnement!

    Digital tilgang er inkludert i alle våre abonnement.

Oracle has developed an open source specification for transmitting tensor data, which the company wants to become a standard for machine learning.

binary, the specification to servers dependencies. memory-mapped and the light while bring data a for querying machine any transmission. network also format provides a protocol on models GraphPipe, efficiency are clients There GraphPipe being simple and from framework. is for deploying learning and intended Called of

includes: It

with Flatbuffers a includes memory Flatbuffer tensors, of similar output input, message set avoiding flatbuffer names. input definitions. Google and - provide protocol an additional during benefit names, of are request that deserialization. definitions to copy buffers, A

serving - models. for Guidelines

learning Examples frameworks. models - from of various serving machine

querying TensorFlow through Clients inside a local Go, a GraphPipe. models a libraries available for library, model for TensorFlow including Python, remote are Java. graph. for served Client Google’s There’s and - for plugin

and message output GraphPipe, tensor about accepts shapes returns and outputs.  of per also model model one inputs name. The request a remote types a and With metadata provides

says … GraphPipe Oracle

IDG News Service