TFRT: A new TensorFlow runtime (TF Dev Summit '20)

preview_player
Показать описание
TFRT is a new runtime for TensorFlow. Leveraging MLIR, it aims to provide a unified, extensible infrastructure layer with best-in-class performance across a wide variety of domain specific hardware. This approach provides efficient use of the multithreaded host CPUs, supports fully asynchronous programming models, and is focused on low-level efficiency.

Speaker:
Mingsheng Hong - Senior Staff Software Engineer

Resources:

event: TensorFlow Dev Summit 2020; re_ty: Publish; product: TensorFlow - General; fullname: Mingsheng Hong;
  1. TensorFlow
  2. type: Conference Talk (Full production);
  3. pr_pr: TensorFlow
  4. purpose: Educate
Рекомендации по теме
TFRT: A new 0:10:10

TFRT: A new TensorFlow runtime (TF Dev Summit '20)

Inside TensorFlow: Functions, 0:46:09

Inside TensorFlow: Functions, not sessions

TensorFlow Runtime by 0:39:54

TensorFlow Runtime by Koan-Sin Tan | COSCUP 2020

What’s new in 0:06:31

What’s new in TensorFlow 2.11

TensorFlow Dev Summit 0:31:12

TensorFlow Dev Summit 2020 Keynote

TensorFlow from the 0:30:15

TensorFlow from the ground up (ML Tech Talks)

TensorFlow Dev Summit 7:50:13

TensorFlow Dev Summit 2020 Livestream

Inside TensorFlow: Building 0:42:50

Inside TensorFlow: Building ML infra

Research with TensorFlow 0:16:01

Research with TensorFlow (TF Dev Summit '20)

MLIR Open Meeting 0:53:46

MLIR Open Meeting 2019-08-01: TensorFlow Graph Representation in MLIR

Inside TensorFlow: TF 0:38:26

Inside TensorFlow: TF Debugging

TensorFlow Fall 2020 0:48:06

TensorFlow Fall 2020 updates: Keynote & what’s new since TF2.2

Tutorial rete convoluzionale 0:10:49

Tutorial rete convoluzionale con Tensorflow!

Mila Agora: Mini 1:50:04

Mila Agora: Mini TensorFlow Summit 9 october 2019.

TensorFlow 2.0 Complete 6:52:08

TensorFlow 2.0 Complete Course - Python Neural Networks for Beginners Tutorial

Audio Recognition using 0:15:39

Audio Recognition using TensorFlow Part-2

MLIR: Accelerating AI 0:10:02

MLIR: Accelerating AI - Chris Lattner and Tatiana Shpeisman

tinyML Talks Kartik 0:37:33

tinyML Talks Kartik Thakore: Building TinyML Applications Using Rune

Ease ML deployments 0:26:00

Ease ML deployments with TensorFlow Serving - Cambridge ML Summit ‘19

The Sky's the 0:14:32

The Sky's the Limit: RE The Next 10 Years of Data Infra on Clouds | Bluesky

Forecasting Stock Returns 0:50:57

Forecasting Stock Returns with TensorFlow, Cloud ML Engine, and Thomson Reuters

Оптимизация и запуск 0:55:14

Оптимизация и запуск моделей TensorFlow с помощью TensorRT (Дмитрий Миронов)...

Guys Likes You 0:00:16

Guys Likes You are Disgrace to the society || tvf meme template

The Power of 0:27:17

The Power of Many: Runtime Environment for Executing Dynamic Heterogeneous Workloads -- Shantenu Jha

Комментарии
Автор

How does XLA fit in all these? By the way, I still can't get XLA to work on my Windows 10 laptop with RTX 2060 GPU. It just won't find ptxas. This must be a bug.

TernaryM
Автор

at 7:48 is current TF optimized with Nvidia's TensorRT?

junkmail