Asplos 23 Session 5c Homunculus Auto Generating Efficient
![asplos 2023 2023 Calendar asplos 2023 2023 Calendar](https://ecocloud.epfl.ch/wp-content/uploads/2022/02/cover.png)
Asplos 2023 2023 Calendar Our evaluations on real world ml applications show that homunculus’s generated models achieve up to 12% better f1 scores compared to hand tuned alternatives, while operating within the resource limits of the underlying targets. This multi faceted nature of in network tools and expertise in ml and hardware is a roadblock for ml to become mainstream in networks, today. we present homunculus, a high level framework that enables network operators to specify their ml requirements in a declarative, rather than imperative way. homunculus takes as input, the training data and.
![юааasplosюабтащ22 Trip Report Sigarch юааasplosюабтащ22 Trip Report Sigarch](https://www.sigarch.org/wp-content/uploads/2022/04/asplos22.png)
юааasplosюабтащ22 Trip Report Sigarch Asplos'23: the 28th international conference on architectural support for programming languages and operating systemssession 5c: machine learningsession cha. Homunculus: auto generating eficient data plane ml pipelines for datacenter networks. in proceedings of the 28th acm interna tional conference on architectural support for programming languages and operating systems, volume 3 (asplos ’23), march 25–29, 2023, vancouver, bc, canada. Homunculus artfiact (asplos '23): the source code and instructions for running the homunculus compiler equipped with the taurus backend. instructions are included for running three different applications. Homunculus docker image (asplos '23): the provided zip (artifact asplos23 docker.zip) file contains a docker image with a pre built homunculus compiler along with all its dependencies. users only need to (1) install docker (and allow non root users to spin up docker containers), (2) download the zipped docker image provided with this archive, and (3) execute the following commands to run the.
![asplos 23 Session 5c Homunculus Auto Generating Efficient Data asplos 23 Session 5c Homunculus Auto Generating Efficient Data](https://ytimg.googleusercontent.com/vi/83Q8nP6J5DE/maxresdefault.jpg?sqp=-oaymwEmCIAKENAF8quKqQMa8AEB-AH-CYAC0AWKAgwIABABGH8gEygfMA8=&rs=AOn4CLDJHaUucdmyMaa5CeU-83YWaXwX7w)
Asplos 23 Session 5c Homunculus Auto Generating Efficient Data Homunculus artfiact (asplos '23): the source code and instructions for running the homunculus compiler equipped with the taurus backend. instructions are included for running three different applications. Homunculus docker image (asplos '23): the provided zip (artifact asplos23 docker.zip) file contains a docker image with a pre built homunculus compiler along with all its dependencies. users only need to (1) install docker (and allow non root users to spin up docker containers), (2) download the zipped docker image provided with this archive, and (3) execute the following commands to run the. Swamy t, zulfiqar a, nardi l, shahbaz m, olukotun k. homunculus: auto generating efficient data plane ml pipelines for datacenter networks. in aamodt tm, jerger ne, swift m, editors, asplos 2023 proceedings of the 28th acm international conference on architectural support for programming languages and operating systems. Homunculus takes as input, the training data and accompanying network constraints, and automatically generates and installs a suitable model onto the underlying switching hardware. it performs model design space exploration, training, and platform code generation as compiler stages, leaving network operators to focus on acquiring high quality.
Comments are closed.