taehoonlee/tensornets

State of TensorRT / XLA support?

Open

#36 opened on Jan 3, 2019

View on GitHub
 (1 comment) (0 reactions) (0 assignees)Python (1,011 stars) (185 forks)batch import
help wanted

Description

Does anyone know if these pre-trained models can be converted to TensorRT engine executables? Or JIT'd by XLA for NVIDIA GPUs?

Contributor guide

Tech stack
pythontensorflow
Domain
machine learningperformance
Issue type
research
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
over 1 week
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
stale
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
unclear
Prerequisites
Basic knowledge of TensorFlowUnderstanding of TensorRT and XLA
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20
Research direction
This issue asks whether pre trained models in tensornets can be converted to TensorRT engines or JIT compiled with XLA. To address this, one would need to investigate the current state of TensorRT and XLA support in TensorFlow, check if any existing conversion scripts exist in the repository, and possibly implement a conversion workflow. The repository's codebase, particularly model definition files under tensornets/models/, would need to be examined for compatibility with TensorRT's requirements such as fixed input sizes and supported operations.