All Projects → tlkh → pycon-sg19-tensorflow-tutorial

tlkh / pycon-sg19-tensorflow-tutorial

Licence: MIT License
PyCon SG 2019 Tutorial: Optimizing TensorFlow Performance

Programming Languages

Jupyter Notebook
11667 projects

Projects that are alternatives of or similar to pycon-sg19-tensorflow-tutorial

vdpau-va-driver-vp9
Experimental VP9 codec support for vdpau-va-driver (NVIDIA VDPAU-VAAPI wrapper) and chromium-vaapi
Stars: ✭ 68 (+161.54%)
Mutual labels:  nvidia
nvtx-rs
A safe Rust FFI binding for the NVIDIA® Tools Extension SDK (NVTX).
Stars: ✭ 28 (+7.69%)
Mutual labels:  nvidia
documentation
Documentation on how to get everything to work and entry point for new users.
Stars: ✭ 72 (+176.92%)
Mutual labels:  nvidia
aws-virtual-gpu-device-plugin
AWS virtual gpu device plugin provides capability to use smaller virtual gpus for your machine learning inference workloads
Stars: ✭ 115 (+342.31%)
Mutual labels:  nvidia
lane detection
Lane detection for the Nvidia Jetson TX2 using OpenCV4Tegra
Stars: ✭ 15 (-42.31%)
Mutual labels:  nvidia
zabbix-nvidia-smi-integration
The Zabbix template for monitoring Nvidia graphics cards.
Stars: ✭ 22 (-15.38%)
Mutual labels:  nvidia
play with tensorrt
Sample projects for TensorRT in C++
Stars: ✭ 39 (+50%)
Mutual labels:  nvidia
nvidialinuxdriversaretrash.github.io
An aggregate void to yell into about Nvidia's terrible Linux drivers!
Stars: ✭ 20 (-23.08%)
Mutual labels:  nvidia
RTX-2080Ti-Vs-GTX-1080Ti-CIFAR-100-Benchmarks
No description or website provided.
Stars: ✭ 16 (-38.46%)
Mutual labels:  nvidia
novideo srgb
Calibrate monitors to sRGB or other color spaces on NVIDIA GPUs, based on EDID data or ICC profiles
Stars: ✭ 222 (+753.85%)
Mutual labels:  nvidia
MinerLamp
Qt GUI for ethminer
Stars: ✭ 40 (+53.85%)
Mutual labels:  nvidia
hashcatbenchmark
Benchmark in Hashcat for diferents GPU's
Stars: ✭ 19 (-26.92%)
Mutual labels:  nvidia
coreos-gpu-installer
Scripts to build and use a container to install GPU drivers on CoreOS Container Linux
Stars: ✭ 21 (-19.23%)
Mutual labels:  nvidia
linux nvidia jetson
Allied Vision CSI-2 camera driver for NVIDIA Jetson Systems. Currently supporting Nano, TX2, AGX Xavier, and Xavier NX. Support for TX2 NX coming soon.
Stars: ✭ 68 (+161.54%)
Mutual labels:  nvidia
cuda-smi
Simple utility to show nVidia GPU memory usage wrt. CUDA device IDs.
Stars: ✭ 36 (+38.46%)
Mutual labels:  nvidia
gpufetch
Simple yet fancy GPU architecture fetching tool
Stars: ✭ 66 (+153.85%)
Mutual labels:  nvidia
ONNX-Runtime-with-TensorRT-and-OpenVINO
Docker scripts for building ONNX Runtime with TensorRT and OpenVINO in manylinux environment
Stars: ✭ 15 (-42.31%)
Mutual labels:  nvidia
JetScan
JetScan : GPU accelerated portable RGB-D reconstruction system
Stars: ✭ 77 (+196.15%)
Mutual labels:  nvidia
peakperf
Achieve peak performance on x86 CPUs and NVIDIA GPUs
Stars: ✭ 33 (+26.92%)
Mutual labels:  nvidia
nvidia gpu exporter
Nvidia GPU exporter for prometheus using nvidia-smi binary
Stars: ✭ 85 (+226.92%)
Mutual labels:  nvidia

PyCon SG 2019 Tutorial: Optimizing TensorFlow Performance

GitHub last commit GitHub

This workshop content covers:

  • a brief introduction to deep learning and TensorFlow 2.0
  • using tf.data and TensorFlow Datasets
  • XLA compiler and Automatic Mixed Precision (AMP)
  • speeding up CNN (ResNet-50) with XLA and AMP
  • speeding up Transformer (BERT) with XLA and AMP

For a quick guide to using Automatic Mixed Precision, check out this TLDR.

Content

Slides are in this Google Drive folder.

Notebooks

Notebook Link Solution
TensorFlow Dataset & tf.data Open In Colab
Pet Classification with TF 2.0 Open In Colab Open In Colab
Transformers with TF 2.0 Open In Colab Open In Colab

For those running the notebooks on the workshop JupyterHub or on your own hardware, you can clone this repository.

git clone https://github.com/NVAITC/pycon-sg19-tensorflow-tutorial

Workshop Information

In-person @ PyCon SG 2019

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].